^ <fr
h ^^
' t
>^.i
t-'^ ■fcrr-*^ V»» ^— ^ ^-
tv i- iv ,?
.> ^ ^ K ^^^>
■' r. ^ ^ t i-
. r r'l' "IT
t. r r fe":
j^:^
r f I?
^.J-.i^,.?^ ..^.,.>..^
4."' f ■
3^ ^-K.^-.X*r,rj^ .^
5^>j/>3:>
j^.j^.j.^
ii' ?^
,"i 'at-
\V. '% I;
|
"J' |
i^ |
1 |
I* f* It ^ ^ ^ iV \i 1 ^ ^ &" |
||
|
* |
"i. |
i |
t '? |
K> {t
- ^' »'v ' »> I'* I-. ^■. .,
i;»^^. iv IX iv 1^ |\ a. N U U li I. I.
i I . Vk: ,-, 1^ 1^ J^ ^
*> i'< Ic- §1.
\i la r
I' I
.^ f ^ ^ f
(^ i^ (i- f<! ^- r# ^ ^ fe t
. i- , : > ^ . ^ I* F^ > ^ i ' ' -
'^ 'v . • , -.>, ^'.> {* f^ ^ f^ r .^■
'• «^ I?- ' '-■ ^ ?1. ?v J.
4
V
ON
AQUATIC CARNIVOBOUS COLEOPTERA
y
OR
DYTISCID^.
BY
DAVID SHARP, M.B.,
HON. MEM. OF THE INSTITUTE OF NEW ZEALAND: MEMBER OF THE ENTOMOLOGICAL SOCIETIES OF LONDON, FRANCE, BERLIN, STETTIN, BELGIUM, AND SWITZERLAND, ETC., ETC.
Plates VII. to XVIII.
[FROJI THE SCIENTIFIC TRANSACTIONS OF THE ROYAL DUBLIN SOCIETY. VOL. 3. SER. IL}
DUBLIN:
PUBLISHED BY THE ROYAL DUBLIN SOCIETY.
PRINTED BY ALEX, THOM & CO., 87, 88, & 89, ABBEY-STREET,
THE QUEEN'S PRINTING OFFICE. FOR HER MAJESTY'S STATIONERY OFFICE.
April, 1882.
TABLE OF CONTENTS
TO
VOLUME II. (SERIES II.)
Pages 1. — Ob.seivatious of NebiiliB aaJ Clusters of Stars made with the Six-foot aucl Three- foot Reflectors at Birr Castle, from the year 1848 up to the year 1878. By the Right Houourable the Earl of Rosse, ll.d., d.c.l., f.r.s. Plates I. to VI. Read Febi-iiary 18th, 1878. Published — Parts 1 and 2, August, 1879 ; Part 3, Juue, 1880, . . . . . . . . . 1 to 178
Appendix, . . . . . . . i to v
2. On Aquatic Carnivorous Coleoptera or Dytiscidse. By David Sharp, m.b., etc.
Plates VII. to XVIIl. Read May 16th, 1881. Published— April, 1882, , 179 to 1003
II.— ON AQUATIC CARNIVOKOUS COLEOPTERA OR DYTISCID.E. Br David Sharp, m.b., Hox. Mem. of the Institute of New Zealand: Member of THE Entomological Societies of London, France, Berlin, Stettin, Belgidm and Switzerland, &c., &c. Plates VII to XVI fl.
[Read May 16tli, 1881.]
1. Prefatory.
Some years ago. I commenced a special study of the Dytiscidoe, with the object of increasing the recorded information about this family of beetles. I had been previously, for a considerable period, specially interested in the family, and when beginning my studies, I hoped particularly that I should be able to improve the very imperfect classification in vogue, and I also wished to know whether a detailed knowledge of the varied structural peculiarities of the species would be consistent with the belief that the present condition of these had been reached by a process of gradual modification or evolution, and whether an intimate acquaintance with the intricate relations existing between tbe diverse components of the family would render credible the hypothesis that these are descended from a few ancestors or even from a single very remote ancestor.
Now that I am offering to the Royal Dublin Society the work that has occupied me for some years, I feel that I must in the first place make an apology for its imperfections and omissions. I have accomplished but little — so little that, in comparison with what I have left undone, I feel it to be almost as nothing. Our knowledge of the earlier stages of the life of Coleoptera, and of their meta- morphoses is very imperfect, and in the case of the aquatic species there are special difficulties in the way of acquiring information of this nature, thus it happens that we know very little of the life histories of the DytiscidjE — so little that it cannot aid at present in the classification of these insects, and I have therefore limited my efforts to producing an arrangement based on the structures of the perfect insect.
Even as thus limited the work is very incomplete ; existing collections though tolerably numerous are very imperfect ; a large number of species are knc'Wii onlv by a single, or by very few individuals, and thus the basis of their taxonomy — ti <• accurate delineation of the characters of morphological species — is still very in- complete, while the important questions of the amount of variation exhibited by the different forms, and of the limits and nature of their distribution on the eaith's surface can be dealt with only in a very inadequate manner.
Moreover, I have made no reference to internal anatomy, for in view of the present position of this department of entomology it is clearly premature to attempt to make a classification founded on both the internal and external structures ; and
TSlAKS. rot. dub. 80C. N.S., VOL. II. 2 B
180 On Aquatic Caraivorous Coleopfera or Dytisc'ahe.
I haw; not endcavoureu to explore the internal anatomy ut" the multitude of species, for to do this in such a way as to be trustworthy for taxouomical purposes would be more than the work of a lifetime.
Thus the memoir I offer to the Society is a mere imperfect sketch of the characters of the species of aquatic carnivorous beetles, and a contribution to a synthetical or natural classification thereof based on tlie external structures of the perfect insects.
I have commenced with a compai'ative sketch of the variations of structure exhibited in the family, and at the same time have touched slightly on its similarities to, and difierences from the terrestrial carnivorous Coleoptera.
The next part is intended to characterize all the species I have examined, and to enable the name, and as far as possible the special characters of anj^ species to be recognized with facilit3\ For those using the work for these purposes it is an advantage to l)e as brief as possible, and I have consequently avoided lengthened descriptions of points common to a considei-able number of species, and have endeavoured to limit the diajT^noses and remarks to a combination of characters sufficient to distinguish a species from the allied ones known to nie. The DytiscidiB however offer in this respect a serious difficulty : owing possibly to the comparative want of variet}^ in the external conditions in which they exist, the species resemble one another much more closely than is the case in other families of beetles ; there is in fact less variety in such details of sculpture and colour as are found in other families to afford important assistance in the recognition of species ; there is too a great monotony of shape or form ; indeed we may say that just as the various ships and boats devised by man have a greater resemblance to one another than have the various carriaijes and machines for travellino- on land, so do the species of water beetles show a less variety in these respects than do the dwellers on land. The discrimination of the species of Dytiscidoe is therefore not an easy task, and recourse must be had for tlie purpose to differences in the minor structural peculiarities. I have found that as I'egards the structure of the feet and legs the distinctions between the sexes vary much from species to species, and I have accordingly given particular attention to these and other secondary sexual characters ; it has been the more advisable to do so because this has hitherto been much neglected, and because the sexes frequently differ so much in their characters that they cannot be recognized as one and the same species unless these distinctions are specially noted.
The analytical tables are merely intended to facilitate the determination of the species by curtailing the number of descriptions to be perused, and with this object the characters used have been arbitrarily selected, and such as are easily observed have been preferred. I have not tabulated the species of each genus, because the attempts I made to draw up such tables satisfied me that in a great number of cases I could not do this in a manner likely to be practically useful.
On Aquatic Carnivorous CoIeo2)tera or Di/tiscid(c. 181
Eleven hundred and seventy-one species are distinguished, in a more or less imperfect manner; and in addition to these there are recorded in the Munich catalogue of Coleoptera or in recently published works some three hundred or more names, a considerable proportion of which probably represent species unknown to me. As these descriptions are extremely scattered in works in various languages, and frequently rare or difficult of access, they are reprinted for the convenience of future workers {vide page 649).
In the nomenclature of the species I have used the system of priority as set forth in my pamphlet on the object and method of zoological nomenclature. I treat the two words forming the name of each species as being practically one, and retain them unaltered, thus for the present abandoning the Linneean idea of making the name of a species indicate its position in classification. The arguments I have used in the paper referred to have not, so far as I am aware, been answered ; and further reflection on the subject has convinced me that the system is not only logically con- sistent, but is likely to be practically useful, and capable of serving as a permanent basis for a system of synthetical classification.
A large and constantly increasing portion of the space of descriptive zoological works is occupied by lists of synonyms : this burden is already so great that it has become a serious question whether zoologists can reduce the extent of these synonyms without danger to the precision and continuity of their w^ork. The plan I have adopted of retaining as the name of the species, the original name, will serve the purpose of limiting future increase of synonyms ; for a very large proportion of those usually attachedto zoological descriptions consists in following the classificatory changes in the name of the species that inevitably result from the unfortunate method of making the name of the species dependent on that of the genus. Another aid is the relegation of pure synonyms (that is such as are not resultant from classification changes but from sheer error or accident) to catalogues published apart from descriptive works, where these synonyms may be recorded once and for all and then done with ; a catalogue should in fact serve as the place of final interment of these unfortunate results of accident, neghgence, or haste. We possess already in the Munich Catalogue of Coleoptera a work in which a large proportion of synonyms are well recorded, and I have considered the existence of this valuable production sufficient reason for omitting the synonymy already recorded therein, and have contented myself with citing in the alphabetical index of this work such names as are necessary to establish a harmony between it and the catalogue in question.
For a similar reason it forms no part of my plan to give a history of the previous and present condition of the taxonomy of the family, nor a list of all the writers who have described species belonging to it : both of these can be gathered from the pages of the Munich Catalogue. The earlier descriptions are scattered in the pages of zoological and entomological works from Linnseus till the time of Aub^ ; special reference is due to the work of the last named author, who in 1S38 published the
3 B 3
1 82 On Aquatic Carnivorous Coleoptera or Dytiscidae.
" Species gcndral des Colcopteres — Hydi-ocanthares et Gyriniens." The descriptions of this justly respected writer are very good, and may continue to be consulted with great advantage in respect to the shape, colour, sculpture, and similar specific characters of the 317 species known to him. A reference to each species described by Aube in the work alluded to, as well as to the species subsequently described by various authors, will be found in the Munich Catalogue. I would also specially call attention to the work of Schiodte (" Genera og Species af Danmarks Eleutherata") published at Copenhagen in 1841, because six of its valuable plates consist of figures of structural details of these water beetles.
The carnivorous water beetles are included in the second volume of the Munich Catalogue which was published in the year 186 8, and in the period that has since elapsed several authors have given us works or notes relating to these creatures. Of such, Crotch " Revision of the Dytiscidje of the United States," (Tr. Amer. Ent. Soc. 1873), Sahlberg "Enumeratio Coleopterorum Carnivororum Fennise (Notiser ur Siillskapets pro Fauna et Flora Fennica Forhandligar, XIV, 1873), Regimbart Etude sur la classification des Dytiscidse (Ann. Soc. Ent. France 1878), and Bedel " Faune des Coleoptercs du bassin de la Seine," (published as a supplement to the Ann. Soc. Ent. France, IS SO,) deserve special notice, inasmuch as their works are of a systematic character, and contain a ([uantity of information arranged in such form that it may be easily consulted and used.
Numerous entomologists and two museums have contributed to the completion of the work, by communicating to me either extensive collections, or a few rare and little known .species. The Musee Royal d'Histoire Naturelle at Brussels, and the Museo Civico di Storia Naturale of Genoa have allowed their collections of this family to remain in my hands for a considerable period, and the Comte Henry de Bonvouloir of Paris has entrusted to me for several years the whole of his collections of Dytiscidoe, comprising the larger part of Dejean's original collection of the family, and many of the specimens actually described by Aube, and for this assistance I offer him grateful acknowledgment.
To Dr. Lcconte, of Philadelpliia, I am greatly indebted for the loan of a series of typical specimens of the species described and named by him in his numerous valuable entomological memoirs. Dr. Horn, of Philadelphia, M. Leon Fairmaire, of Paris, The Rev. A. Matthews, of Gumley, Leicestershire, Prof Sahlberg, of Helsingfors, and Herr Ernst Wehncke, of Harburg on Elbe, have all loaned nie specimens of which I had need and for doing this I heartily thank them. I will also mention here that I adopted the plan of determining collections and specimens sent to n)e by means of a number attached to the species instead of a name ; the number used by me for this purpose will be found attached to each species in the prestnt work, and is placed after the description of the habitat in each case.
The concluding part of this memoir is intended to be a contribution towards a natural classification of the species previously characterized. This classification I
On Aquatic Caniivorous Cohoptera or DytisciJcje. 183
have attempted to make on a purely synthetical method. Linnaeus and the more im- mediate of his followers made their classifications by a deductive or analytical method. Natural objects were taken for granted as forming three separate kingdoms, animal, vegetable, and mineral — and then each of these hypothetical or axiomatic kingdoms was divided and subdivided until the genus and species were reached and defined. This method — most useful in the early stages of scientific develop- ment— is being giadually replaced by an inductive or synthetic system. In other ■words species are studied and defined and then treated as units whose relations to one another may be exhaustively considered. This method has been clearly recognized by fTuxley as will be seen by consulting his paragraph on morphological groups ("Man. Anat. Invertebrated Animals," p. 17) : and it is quite evident that only by this synthetical method can we hope to ascertain whether genera and the more complex aggregates have any real existence in nature, or whether they must remain — as they undoubtedly are at present — merely conventional arrangements. On tlie Linntean method it may be said with perfect truth that the genera were made by the system, whereas it is clear that on a natural science method the system must be made and built up in accordance with the actual structures ascertained detail by detail. If there be an order in Nature our zoological systems must be made in accordance therewith, and representative thereof; if there be no such order, we may then revert to the Linnafjan method, for this is well adapted for purposes of artificial arrangement. At present I scarcely think there is any conscientious naturalist who after a careful review of facts would say whether there are or not in Nature, independent of classifications, such things as genera. I believe, if we limit our view to the creatures coexisting at the present moment, no naturalist could be found who would venture to deny the^existence of species as real and objective. It is in fact perfectly clear that the hosts of individuals livingaround us are arranged in clusters or groups, isolated from other clusters or groups ; and although there may be doubts as to the actual number of such clusters — and doubts as to whether certain masses of individuals form two isolated clusters or only one, yet no practical naturalist will be found who will deny the reality of the existence and isolation of such clusters, and it is these we call species. The nature of the connection between the individuals of these clusters, and the kind of isolation existing between them are most difficult questions, but their discussion has been commenced by Lamarck, Darwin and others; and these problems are now recognized as legitimate subjects for scientific investiga- tion, although perhaps but few perceive their excessively complicated and difficult nature.
If, however, it be granted that species have a real existence, and if their characters have been well ascertained, it is clear that we may then deal with them as units to be classified and arranged, preliminary to considering the question whether these species are, like the individuals of which they are composed, arranged in naturally isolated clusters. This is evidently a much more complex question to deal winh
1S4 On Aquatic Carnivorous Coleoptcro. or Dtjtiscidie.
than that of species; l)ut il" we suppo.se it to be answered in an affirmative sense^ we may then treat these complex aggregates as themselves units, and study them in the same way as we have supposed to be already done in the case of genera and species. Looking at taxonomy in this aspect, species are the units of the classifier, genera are aggregates of the fir.st degree, groups of genera aggregates of the second degree, and so on.
Whatever view be taken as to tlie nature of genera, it will at least be admitted that there does not prevail among naturalists at present so great concord as to their limits, as that which we find to exist about species. It is undoubtedly the case that an attempt to define genera in a thoroughly natural manner is much more difficult than the study of species ; while an additional obstacle is thrown in our way by the fact that we are not yet acquainted with all existing species, so that any da}^ by the discovery of a new and intei'mediate form two apparently distinct genera may be connected together.
This process of forming genera by synthesis — by the accord existing between species as to the whole of their characters other than specific — has notwithstanding its difficulties jirogressed in a more or less recognized manner, and is no doubt destined to su])plant completel}' the process of ai'tificial classification; and there can I think be but little doubt that the perfection of the miethod will be, for a while at any rate, accompanifd liy an increase in the number of primary aggregates recognized by zoologists. The enormous increase that has already occurred in the number of genera has occasioned much discussion and given rise to considerable complaint against those who make new genera ; and in point of fact the history of systematic zoology pi-esents us with a picture of constant protest by the older naturalists against the uuiltiplication of genera by their younger fellows. And it nuist be admitted that the increase has indeed been enormous ; Linnseus only admitted 354 genera as composing the whole animal kingdom (" Systema NaturiB," Ed. xii) ; the nuiuber at jiresent recognized could not be readily ascertained, but it is something enoimous ; there are I believe about ten thousand genera recognized merely among the beetles, and I anticipate that this number will increase to forty or fifty thousand. I do not myself feel alarmed at this extraordinary nuiltiplication, but there is one poini connected with it that is certainly to be regretted ; it is the corresponding increase in the number of generic names ; the enormous growth in this I'espect has already brought us to the unfortunate pass that few of us, even tliough we are specialists, recognize, much less understand, the generic names even of the depart- ment to M'hich we have given particular attention ; and thus the names of genera come to have but little value. Seeing then how uncertain are the limits of genera nt present, and iiow doubtful it is what kind of relation the genera we now adopt may bear to those of the future when systematic zoology shall have assumed a more periect form, we may Avell consider whether some method cannot be devised to limit the incnnise of generic names, or at any rate to render their recognition a
(Jn. Aquid'tr Carnivorous Coleoptera or Di/ti-!cida\ 185
matter uf subordinate importance. And it seems to uie that a muiierioal linear arrangement might here be of assistance. If genera were numbered consecutively, ^nd the number habitually quoted, we should recognize tlie position of the genus much more easily by means of the number than we do by means of a name; for instance if we have the knowledge that there are about 700 genera of Longieorns and have an idea of the method on whicli they are arranged, and if we knew that in a numerical arrangement of the Coleoptera the Longieorns begin at 3,000, and run on to 3,700, it is clear that we shall at once recognize any number between the tuo just named as indicating a genus of Longieorns : and further that we shall be able to recall in a more or less vague manner, what part of the Longicorn family it belongs to ; whether it comes near the beginning or end of the family and consequently «ome ally with which we are specially familiar. There is a further advantage in a numerical expression of classification, inasmuch as it gives us a prospect of being -able in the futui-e to indicate in an exact manner, by means of a numerical formula the true systematic value of any aggregate we are required to deal with, and thus to ascertain for instance the true difference between the faunfe of two countries we might wish to compare in this respect. A numerical system applied in the way I have indicated, can however express only a linear arrangement, and as this can never be more than an approximation to a natural classification, it is clear that a simple numerical system can never do more than express feebly the complex, relations existing between animals ; moreover, it could only be of full service when applied systematically to all the animal kingdom ; and I have accordingly adopted it here as an adjunct of my classification without attaching to it any considerable present importance ; but still it seems possible that it is in the direction of numerical expression of classification that we may hope for real improvement.
In the second synthesis the genera are treated as units to be arranged into groups ; and in the third synthesis these groups become the units of which tribes -are formed ; and this metliod of prooeiKling from the simpler to the more complex aggregates is carried on till ultimatel}^ such questions would come to be considered -as whether there really exists an isolation or division of organic beings into vegetable and animal kingdoms. In the present memoir this method of classifica- tion is carried only to the extent of a synthesis of the fifth degree.
I have already mentioned that during the progress of the work I have had my -attention frequently directed to those problems as to evolution and descent in which naturalists at present are so greatly interested, and the points I have more particularly kept in view were these two. First, are the structures of iiiese creatures of such a kind as to make one believe they may have arisen by gradual modification of a precedent different structure ? and second, are the various re semblances, or affinities as they are called, among the different species such as to make it appear probable they are the results of genetic community of descent
186 On Aquatic Carnivorous Coleoptera or Dytiscidoe.
from one ancestor or a few ancestors. The first of these questions I can answer decidedly in the affirmative. That these water beetles have arrived at their present condition by a process of gradual modification or evolution, seems possible and con- sistent with their structures ; indeed I may go farther and add that there are some points of their structure which are not comprehensible on any other hypothesis known to mc.
On the second part of the problem I proposed to myself, viz, whether the great number of species of Dytiscidae now existing have probably descended from one or a few ancestors, 1 have come to a decidedly negative conclusion. After a care- ful study of the various affinities and points of structure, I cannot consider them as indicating genetic community, and I have come to think that it is more probable that each species has been evoluted along a distinct and separate line of descent. Thus I fancy I see in tliis mass of twelve hundred species, not a development from one ancestor but the results of twelve hundred lines of development. The numerous cross-7-elations between the various aggregates, and the points of re- semblance between different species seem when first examined almost irresistibly to suggest that they may be accounted for by assuming descent from a common ancestor, but more careful study instead of rendering this more probable has always had tlie opposite result. One conclusion, I think, I can state almost positively; it is this, that whatever may prove to be the connexion between existing and extinct morphological forms, there is no relationship of an ancestral or genetic kind to be traced between actually existing species. This result although negative is not without significance, for among these twelve hundred species there are many in a later or higher stage of evolution than others, and yet in no case have I been able to consider that a lower existing form is ancestral to a higher existing form ; the theory of descent from a few ancestors would however lead us to suppose that, in some cases at any rate, parental species and descendant species should for a time co-exist. In the various syntheses forming the third part of this memoir there will be found some sketches illustrating the kind of reasoning that has brought me to these conclusions. Although quite inclined to agree with Huxley's remark (" Manual of the Anatomy of Invertebrated Animals," p. 4) " that the growing tendency to mix up setiological speculations Avith morphological generalisations will, if unchecked, throw biology into confusion," yet 1 think it must be admitted that if there is ta be any expression of opinion on aetiology, it is well that it should be plueed in proximity with the observations on which it is founded, for only in such ca-e can its true value be appreciated, and I hope in the present instance it will be found that the few remarks I have made on these points in no way detract from the value of the observations with which they are associated. I will ask also per- mission to make now some brief remarks on these setiological problems, my object being not to advocate any particular theory, but rather to reiterate the extremely difficult, nature of these questions, and specially to point out that even if the theory
On Aquatic Carnivorous Coleoptera or Dytiscidce. 187
of evolution be true, community of descent is not rendered probable thereby. I have already stated that the existence of taxonoraical seiies does not really lend any support to the theory of a few ancestors, and I would now add to this that similarity of structure as to some particular point also fails to establish the pro- bability of community of descent ; for if one hypothetical ancestor can have developed a point of structure, equally can two or more have done so.
Semper has recently pointed out that " every character which can be regarded as a true sign of large groups of animal forms, may be ultimately traced to the stage at which it first appeared, and where it was a character of ada23tation " ("Animal Life," notes p. 407). This is perfectly correct, and to it may be added the fact that if we look at some character that is now ^Drobably in process of adaptation, we find that the adaptation is going on not in one favoured species, but in a number of allied species. For instance, it is the rule in the Dytiscidte that the meso- and metasterna are connected together in the central line of the body ; a considerable number of the group Hydroporini form however an exception to this rule, and one of these genera, Deronectes, differs only from Hydroporus by the fact that this connection is wanting in the former while it has been attained in the latter genus ; but a study of the species of Deronectes seems to show that the connection m question although not at present existing is j^i'obably being gained by many of the species if not actually by all of them. Now if a structure be acquired simultaneously by a number of distinct species, it is clear that similaiity of structure does not indicate community of descent. Again if I am right in supposing the species of Deronectes are acquiring this structure by the want of which they are solely dis- tinguished from Hydroporus, it is plain that these two distinct genera are, so far as can be seen, in process of becoming one : the real difference between them is in fact one of time — Hydroporus has gained a particular structure before Deronectes has done so. Such facts apjjear to me gravely opposed to the a priori probability of descent from a few ancestors. But if the resemblances between animals do not justify the theory, there remains on the other hand the important fact that the isolation of species from one another is gravely opposed to its probability. Huxley has pointed out in his essay On the origin of .species (" Lay Sermons, Addresses and Reviews"), that the term species expresses two different sets of facts ; first, a set of morphological facts, or the agreement of series of individuals in points of structure ; and second, the physiological inductions that animals consist of groups of individuals that are fertile inter se, but who do not produce fertile offspring when crossed with members of other groups ; and in the same essay he has stated that when we look at the facts from the point of view of the morphological agreement between indi- viduals then the theory of community of descent is possible or probable, while if we bear in mind the jihysiological isolation of species then the theory is unsatisfactory. The arguments of Huxley in this essay appear to me very good, except that as I have already pointed out the morphological facts do not when carefully considered sujjport
TRANS. ROY. DL'B. SOC . N.S., VOL. II. 2 C
188 On Aquatic Carnivorous Colcoptcra or Dytiscidce.
the hypothesis to any great extent. I think his argument might be carried a step farther, and it might be pointed out that whereas the acceptation of the theory of evohition in no way diminishes the importance of the physiological differences between species, it does on the other hand very much detract from the value of the morphological agreement between them ; at any rate it is I think clear that the physiological distinctions between different species may be real and permanent even though the structural characters by which they are represented in our classifications shall prove to have been fluctuating or temporary.
We know from paleontology that the totality of the organisms of the present day is very different from what it was formerly as regards the morphological structure ; but we di not finrl traces of any facts that would lead us to believe that the earlier organisms were not isolated into separate species just as the present ones are. While the excessive resemblance between the ova of different animals coupled with the fact that these similar ova develope into totally distinct organisms, leads us to believe that distinctions of a most profound character may have been present in the earlier and simpler organisms even though the imperfection of our means of investi- gation would not allow us to assert this from the study of the organisms themselves. Thus it seems to be a perfectly credible theory that the isolation of species from one another is indicative of deej) seated distinctionsthatmayhaveexistedpriorto the more conspicuousmorphological differences ; and itappears probable that though the species of a late geological epoch may be the descendants of differently structured animals of a former epoch yet the physiological distinctions may have been more permanent, or in other words that the lines of descent have been perfectly distinct. It will at any rate be incumbent on those who advocate community of descent to show how and when the physiological distinctions became established, and at present, so far as I know, there is no evidence of the beginning of such distinctions ; we are not aware of any process by which a group of individuals fertile inter se, becomes divided into two or more distinct, mutually infei'tile groups. But those who maintain the community of descent tacitly assert that such a jirocess has been the order of Nature.
There seems good reason for supposing that the physiological distinctions between species are correlative with molecular distinctions that we do not at present comprehend ; and if so we may hope that the advance of the science of pure molecular physics will help to solve these problems. But if we are not content to wait patiently for such a period, and at present allow our imaginations to attempt to penetrate very far back into the darkness of the past, we may well ask those who maintain the descent of organisms from one or a few ancestors, what reason have they for supposing — what satisfaction can there be in believing — that only one or two or a few primitive organisms or germs existed ?
Surely, if the passage from matter in a state of inorganic arrangement to a state of more complex organic arrangement took place naturally — under the influence of
On Aquatic Carnivorous Coleoptera or Dijtiscidce. 189
far reaching and enduring law — the transition must have been long enduring in time and wide spread in space. Can we believe it possible that universal law should have resulted in the production of a single inconceivably minute portion of organic matter, from which all the enormous aggregate of organisms now around us has been produced in its almost unthinkable variety by a process of simple reproduction? No, I think that whatever the mode or modes may have been by which inorganic matter became organic, it will surely prove more credible that the passage extended over a vast period of time, and occurred in numerous places, and was not limited to one instant of time and a single pin point of space. It we say this we admit that it is possible that there have existed distinctions in the individual masses of oi"ganic beings from their very beginnings, and it is quite conceivable that in these aboriginal molecular diflerences may have originated the present physiological species distinctions, — that these are in tact absolute and direct continuations of primeval molecular ditferences of constitution ; and it is credible that from such a myriad beginning, through the enormous ages of the woi'ld there have developed the vast multitudes of species of plants and animals amongst which we live. If anything like this has been true, we need not necessarily adopt any theory of community of descent, but we may believe that each species is a distinct record of the past conditions in which it has existed, and that resemblance in structure of two different species is the result of similar growth under similar conditions.
I shall not at present allude farther to the difficulties that surround the theory of community of descent, but I may remark that even if it should prove that we must abandon the hope of tracing the pedigree of all creatures back to a single organism, this in nowise detracts from the importance of biology. We are not called on to abandon the attempt to understand the relations between existing and extinct morphological forms, species by species, and to trace the road by which existing structures have become what we see them. The theory of evolution is in no way connected with the hypothesis of common descent ; and by means of the perfected acquaintance with the structures of existing organisms we shall attain, and of the detailed knowledge we shall acquire of the special modifications that have taken place in myriad separate lines of descent on various parts of the earth's surface, we may well hope that we shall be able to read slowly but truly the great history of Nature.
Bates has an admirable remark — (" Naturalist on the Amazons," Vol. II, p. 34.5) speaking of the local variations in the patterns on the wings of butterflies, he has said, " On these expanded membranes Nature writes, as on a tablet, the story of the modifications of species, so truly do all changes of the organization register themselves thereon. " By the evolutionist this sensitiveness thus truly claimed for the butterfly's wing, may logically be asserted to have always existed in all the structures of every species of the organic world. Every individual is a mass of
2 0 2
190 On Aquatic Carnivorous Coleoptera or Dijtiscidie.
structures, each of which is a record of past action (of the discharge of function under curtain conditions) and thus in the structure of every organic being there is written a portion of the histoiy of the world. Just as the coarser facts of that great history are written in the forms of the hills and valleys, the shores and the abysses, so are the minute details recorded in the forms and structures of organic beino-s. To read that history and understand it, is the task of many generations, it is but recently we have learned that such a history exists to be read, and we have as yet scarcely mastered its alphabet.
II. Comparative Review of the Structures of the Family.*
Size. — The Dytiscidaj are of very variable size, Notomicrus brevicornis being only 1 mm. in length, while Megadytes ducalis attains nearly 50 mm. The most minute forms are found among the Noterides, Hydrovatini, Bidessini, and Hydro- porides ; the most massive amongst the Cybistrini and Dytiscini. The Noterides, Laccophilini, and Hydroporides, do not comprise within their limits any individuals so large as the smallest of the Cybistrini and Dytiscini ; and in respect of this point the Colymbetides and Hydaticides stand intermediate between the small insects and the large ones. The linear classification I have adopted agrees, as a whole, in a remarkable manner with the development in size of the individuals composing the groups : all the small forms come in the earlier half of the arrangement, while the massive Cybistrini terminate it. It may be truly said, in a generalized manner that the hiffher forms consist of individuals of much larsrer size than do the lower forms ; no doubt many exceptions to this generalization may be pointed out ; and much variety exists in respect of size within the limits of a single genus ; for instance, in Cy bister we have the little C. dehaani only 13 mm. long, while C owas attains 40 mm. of length.
In Form the Dytiscidse show comparatively but little variety, in fact there is probably no other family of Coleoptera of equal extent so uniform in this respect. The outline is oval or oblong-oval, the grea,test breadth being usually attained a little behind the middle of the length of the body : considerable differences are found in the amount of elongation, sometimes the oval is so short that the outline approaches to the circular form, as is the case in many species of Hydrovatus and in Colpius, while in the case of Dyiiscus cicurus. Fab. (No. 95G, Ehantus,) the length is nearly 2^ times the greatest breadth ; all the intermediate grades of elongation occur. The extremely ^hort forms are found only among the smaller species. Sometimes the anterior extrem- lity of the body is broader than the apical, as in Matus, but more usually the reverse is the case, and in Thermouectes the hinder portion of the body is very much broader than the anterior half ; in certain rare cases the extremity of the body and the wing-
* This sketch should form part of the fifth synthesis, but is placed in this position in conformity with the usual custom.
On Aquatic Carnivorous Coleoptera or Djtiscidce. 191
-cases are so acuminate as to be spiiiose (Celina, Methles, some Hydrovati) ; this occurs only in species whose individuals are of small size, and in'all such cases the ■swimming legs are very feeble. The most remarkable peculiarity in the outline is its continuity, or freedom from interruption at the junction of the base of the prnthorax with the wing-cases : this continuity is frequently so perfect that no break in the regularity of the form can be detected at the spot indicated ; this is the case with all the higher forms, and the best swimmers. But there are in the family some exceptions in which there is a great break in the outline at this spot, the base of the thorax being much narrower than the base of the elytra (Vatellini, Tyndallhydrus, Andex, Anisomera) ; this discontinuity is always associated with a peculiarity in the sternal pieces ; in its extreme cases such as in the Vatellini, Andex, and Tyndallhydrus, the prosternnl process does not reach the mesosternal process, while in those cases where the discontinuity is less marked, it will be found that although the prosternal process ma}' attain the metasternum, the mesosternura is less completely vertical in its direction than usual ; this will be found to be the case in Agabus cephalotes and in Dytiscus dorsalis (No. 630, Hydroporus). In the genus Deronectes it frequently happens that there is a considerable difference between the .sexes of one species in respect of the continuity of the outline of the thorax and wing-cases and it is always in the female sex that the continuity is the more broken. {See gen. Deronectes, No. 454 Hydroporus lareynei. No. 455 Hydroporus opatrinus, No. 472 Dytiscus depressus, &c.)
The group Noterides presents a constant difference in outline from the other Dytiscidae inasmuch as the greatest width is in front of the middle of the body instead of behind it, and the species of this group have consequently the posterior portion of the body narrower than the other Dytiscidse, so that in them this part is less broad than the anterior half, while, as we have just seen, in most other species of the family the reverse is the case. As regards the convexity of the body the Dytiscidae show a great deal of variation, they are either subdepressed or little -convex, or they may be very convex ; the flattest or most depressed forms are found in the tribes Colyuibetides and Hydaticides, especially in the genera Copelatus, Lacconectus, Platynectes, and (a few) Agabus, while the most convex species are all small insects, of short form, such as Colpius and Suphis among the Noterides, most of the species of the groups Hydrovatini and Hyphidrini, and many of the group Bidessini ; in the Hydroporini some of the genera consist of convex forms, as in Hyphoporus, Herophydrus, Coelambus and Chostonectes ; while Deronectes and Hydroporus comprise forms which as a rule are much more depressed. The larger .and more powerful forms are usually of moderate convexity. When the convexity is very great it is as a rule confined chiefly to the under surface ; this is shown in a very marked manner by Pach3^drus and Hyphydrus ; in the convex Hydrovatini the convexity is obtained by about equal curvature of both the upper and lower surfaces; the Noterides display another peculiarity in this respect inasmuch as being
192 On Aquatic Carnivorous Coleoptera or DytiscidcE.
always convex insects, the convexity is chiefly exhibited by the upper surface, the lower surface being unusually flat ; this peculiarity is however reproduced in the Colymbetides by many species of the genus Th'bius. It will thus be seen that as regards foiTo the Noterides are more peculiar and isolated than any of the other Dytiscidfe they being peculiar in their outlines both in the perpendicular and horizontal directions. Although peculiarities of form are usually to a large extent constant in allied species and genera and are even found to be characteristic in some of the still larger aggregates, yet exceptions are of frequent occurrence ; thus although a certain peculiarity of form is very characteristic of the genus Ilybius, yet there are certain species of Ilybius in which this peculiarity is absent, and which would therefore at first sight be considered to belong to the genus Agabus. In other cases variation in facies goes to a still greater extent ; thus the species of Lacconectus resemble Laccophili in size, form, and colour (although they have no approximation to that genus in any natural classification) and would scarcely be considered to belong to the Colymbetides by any one who examined them for the first time ; so too in many species of the genus Copelatus we find in their appearance but little of the Colymbetides, they more resemble the Hydroporides ; while certain Noterides although they have no aflSnity with the Hydroporides have been actually up to the present time always classified with them, and indeed placed in the genus Hydroporus.
In Colour the Dytiscidie show much less variety than do the beetles dwelling on land ; brilliancy of colour is indeed quite absent. The usual colour is black, or yellow, or a combination of the two, or else a colour intermediate as it were between the two ; faint greenish, olivaceous, and brassy tints appear in the Colymbetides, Dytis- cini and Cybistrini ; the gayest and most variegate colouration is found in the tribe Hydaticides, where a mixture of black and yellow, in conjunction with a very shining polished surface gives rise to a very agreeable appearance : the colouration of these Hydaticides seems however to be in many cases so very inconstant that it can scarcely be of much service in defining the species, or in assisting in their recognition. In the ColjTiibetini, especially in the genus Rhantus, the el}i;ra are frequently of a yellow colour, which is however greatly obscured by a vast number of minute black specks or dots, these become in some species so dense and numerous that the wing- cases appear nearly or quite black, though frequently their speckled nature may be detected on a closer examination, especially when the lateral portions are looked at, the specks being there always less dense than near the suture. Several species of Hydaticus have a similar system of colouration and indeed so much resemble these Colymbetini that the species of the two widely separated genera are still much con- founded together, even by entomologists of good repute. In the genus Hyphydrus and in many Hydroporini {e.g., Coelambus and Necterosoma) the prevalent colour is yellow or yellowish with black marks on the upper surface forming a very irregular pattern ; the black marks being usually more or less elongate longitudinal lines, which
On Aquatic Carnivorous Coleoptera or DijtiscidcB. 193
frequently run into one another laterally, so as not to be readily distinguished. Many species of Deronectes are also variegate with black and yellow, but the colouration is less distinctly of a linear character than it is in the insects just referred to. The Laccophilini vary much in colour, and are usually pallid ; frequently they are marbled with darker tints in a very indefinite manner, but some species {e.g., Nos. 90 & 97) become definitely and prettily marked : a very curious colouration is found in several species of Laccophilus (No. 156 and allied species) consisting of numerous very waved longitudinal dark lines, such as I have not seen in any other beetles. The most pallid species of the family are the two Eretes; E. australis is of an entirely pale colour, except that the wing-cases are covered with numerous rather large punctures, each of which is black; the second species of the genus (Dytiscus sticticus L.) has the same peculiarity, but has in addition a black fascia on the wing-cases, and frequently also another on the prothorax. The colouration of the wing-cases in Acilius resembles that in Eretes, except that the punctures are not so large and definite and the black colour is not limited to a single puncture, but connects two, three or more punctures together, giving rise to an irregularly speckled appearance approximating to that which I have already described as existing in the genus Pihantus : the irregular black fascia on the wings of Dytiscus sticticus just alluded to is produced in a similar manner. An entirely black surface is quite frequent in the family, and is found iu Noterides (Hydrocanthus Australasioe and H. funebris) in Hydrovatus, Hydroporus, and in numerous species of Agabus, iu Copelatus, Ilybius, Meladema, Cybister, &c. One of the most marked features of the colouration of the DytiscidfB is the fact that the lateral margins of the thorax and wing-cases are very frequently of paler colour than the rest of the upper surface, so that the body bears a more or less definite band of pale colour defining the outline of the upper surface; this yellow band finds its greatest development in the genera Dytiscus and Cybister, where it very frequently forms a perfectly definite lateral stripe of yellow colour ; in other genera, however, the stripe is not so distinct and definite {e.g., Ilybius fenestratus) and even in Cybister there are some species where the lateral margin is only indefinitely paler. In Ilybius and numerous species of Agabi which have a very dark coloured upper surface, a minute dot or dash of pale colour (or both of these) exists on the wing-cases, near the apex and lateral margin.
In the Hydaticides the head and thorax bear with great frequency transverse marks of pale and dark colour on the head and prothorax and in numerous species of Dytiscus the whole of the prothorax is definitely margined with yellow ; and this is found again in two species of Cybister (Nos. 1,154 and 1,155).
Although the colour of the body is so frequently black, the antennae nearly always remain pale; and it is the rule in the family that the palpi and legs, as well as the antennae are paler than the rest of the body ; the swimming legs however -are very frequently darker in colour than the other legs.
Another peculiarity of frequent occurrence in the family, is the existence of two
194 Oa Aquatic Carnii'orous Coleoptera or Dytiscidw.
indefinite spots of pale colour on the vertex of the head. The upper and under- surfaces of an individual are frequently very different in colour, the one being often pallid while the other is entirely or in great part dark {e.g., Dytiscus rceselii, Cybister, No. 1,169) : it is, I believe, always the case, when such discrepancy pre- vails, that the upper surface is darker than the under, and the latter is never much variegate as the upper surface sometimes is. As a rule the colouration of the two sexes of one species is nearly or quite similar, but to this there are a few singular and interesting exceptions ; the colour of the ventral segments in Hydroporus bistri- gatus (Bidessus, No. 299) is red in the female, black in the male ; in other species of Bidessus the same fact is repeated but not in so conspicuous a manner ; and we find it again among the Hydroporini in Coelambus (flaviventris, pallidulus, &c.) and in Hydroporus (Dytiscus halensis, No. 528) and again among the Colymbetides in Lancetes (Colymbetes prsemorsus. No. 916). In all these cases it is the males that have the ventral segments dark ; the repetition of this character in a few widely separated members of the family is a very interesting fact and should engage the attention of those who are interested in the questions of sexual colouration and selection. Indeed the whole family of the Dytiscidae is one that seems peculiarly adapted for throwing light on the questions of development and correlations of colour in the insecta, owing to the comparative simplicity of the facts ; most of the other families of insects show such a variety of colour and pattern that these are quite bewildering to any one who wishes to comprehend their meaning ; but the facts in the Dytiscida3 are less complex, and appear to be capable of being generalized in such a manner as to lead us to hope we may some day understand them.
In the Sculpture of the surface we find that the Dytiscidae show less vaiiety than the other families of Coleoptera : to this rule there is however one most interesting exception in the fact that difference — and frequently extreme difference— in the sculpture of the sexes is of very common occurrence. Leaving this for a while out of the question, we may note that in the larger and in the more highly developed forms the surface is remarkably smooth and polished, and punctuation is nearly absent. In these smooth species there exist, however, with remarkable constancy three longitudinal series of fine punctures on each wing-case ; the series are about equidistant from one another and from the suture and outer margin, but frequently the external one of the three is much more indistinct than the others ; the persistence of these punctures throughout the family is remarkable, and seems to indicate that they have some considerable functional importance. In the coarsely sculptured and punctate species, they are more difficult to trace than they are in the smoother species, not only on account of the roughness of the surface, but probably also because they are really less developed. Each of the })unctures in these series bears a very fine hair, and sometimes the series can only be detected by means of these hairs, which can be distinguished as forming a series, although the punctures them-
On Aquatic Carnivorous Coleoptera or Dytisciclce. 195
selves are lost amongst the general punctuation of the surface : in some cases (Colpius inflatus e.g.) I fail however to find any trace of their existence, and in. other cases {vide Laccophilus) they are excessively fine and indistinct, and as a rule are most regular and highly developed in large and smooth species such as we find in the genus Cybister ; as already remarked the outer series is more irregular and less persistent than the others, the sutural series being the most permanent.
In the great tribe Hydroporides, the rule is that the upper surface of the body bears a true punctuation such as exists in many other families of Coleoptera, while on the other hand in the Colymbetides and Hydaticides, the opposite of this is the case — punctate species being exceptional.
When present, the punctuation is of various degrees of fineness and coarseness, of density and sparsity, and is more usually seen on the elytra than on any other parts ; in many of the genera of Hydroporides, however, the hind coxae are very persistently punctate. The most remarkable punctuation occurs in the genus Pachydi-us, where certain species have a deep coarse punctuation on the wing-cases, the punctures moreover being more or less elongate in a peculiar manner. In many species of the genus Hyphydrus the elytra have a double punctuation, very fine punctures being mixed with considerably coarser ones.
In the Agabini punctuation is nearly absent, the surface being either smooth, or — and this is the more frequent — covered with a very fine reticulation, formed by fine scratches placed so as to form meshes of irregular shape, and varying much from species to species. In Copelatus the sculpture is peculiar, and in its highly developed form consists of numei'ous longitudinal grooves or stri;B on the wing-cases arranged in a very regular manner. In the Colymbetides the most peculiar sculpture of the family is found : in Colymbetes the wing-cases are marked with transverse scratches, striae or grooves, placed very near to one another, and varying greatly from species to species. In Meladema {vide M. coriacea No. 978) a still more remarkable sculpture exists ; the elytra bear a large number of crescentic marks, which at the base are readily distinguished as such, but towards the extremity of the body become somewhat changed in shape and are very densely placed, so that an appearance is caused somewhat as if the surface were covered with overlapping scales.
In the Hydaticides the surface is usually smooth, but in Eretes the wing-cases are marked with rather large quite round punctures, and in Acilius both the upper and under surfaces of the body are much punctate. In some species of Noterus there exists a peculiar punctuattion consisting of large round isolated punctures placed chiefly on the hinder part of the wing-cases ; and in Synchortus there is a somewhat similar sculpture, except that the punctures have their front margins somewhat raised giving rise to a peculiar rough appearance.
It is worthy of notice that when punctuation is absent from the greater part of
TRANS. ROT. DUB. SOC, N.S., VOL. II. 2 D
196 On Aquatic Carnivorous Coleoptera or Di/tiscido:.
the upper surface, it still often exists to a greater or less extent on the apical portion of the wiug-cases [vide gen. Dytiscus, numerous species) ; and in the case of many species where the sculpture is peculiar, it frequently loses its peculiar character to a greater or less extent on this part and approximates on it more or less to ordinary punctuation.
The anomalous Amphizoa presents us with the only case among the Dytiscidee where sculpture approaching to what is so usual in the Carabidse exists. In the latter family the elytra very frequently bear each eight or nine rows of punctures or striae or punctate striae, and in Amphizoa something of the same sort is found, but the sculpture in this case is without the regularity and definiteness it possesses in the Carabidae. Elevated costae or ribs on the wing-cases are very rare in the Dytiscidae, but occur in Darwinhydrus in conjunction with a punctate surface. In many Hydroporini the surface bears an extremely fine, short, depressed pubescence, but this is an exception to the condition usually prevalent.
It is worthy of note that in nearly all the cases in this family where we find a peculiar sculpture present and well developed, we may see in allied species such sculpture present in a more rudimentary manner, and thus, as it were, indicating the steps by which it has been developed : this is well exemplified in the genus Copelatus (for notes on the sculpture of which, see page 201).
The most remarkable fact in the sculpture of the Dytiscidae is the sexual disparity found in numerous species. This sexual disparity when present is of such varied character that I can here only allude to a few instances. In the first series — Dy tisci Fragmentati — no instances of well marked difference in the sculpture of the sexes have been recorded, but it is possible that the peculiar punctuation observed in certain species of Synch ortus {vide Hydrocanthus asperatus No. 18) may prove to be confined to the female sex ; the differences existing in the sculpture of the prosternal process in the two sexes of some sjiecies of Hydrocanthus is accompanied by a change of form in the part, and is no doubt a phenomenon of a quite different category to the remarkable distinctions in the sculpture of the upper surface found in the Dytisci Complicati.
In many species of Hydrovatus the distinct punctuation of the female is the same as in the male, but yet the former sex has the surface rendered dull (coriaceous) by a very fine additional sculpture.
In Hyphydrus we find that the females have the punctuation of the upper sur- face very obsolete in comparison with the males, but, nevertheless, in the former sex there is a development of an excessively fine sculpture which renders the surface quite dull : in this genus we find that certain species have two forms of the female, one of them resembling the male in sculpture while the other is dissimilar in the manner just described ; and this phenomenon of dimorphism in the females continues to be of very frequent occurrence in the Dytisci Complicati. Passing to the group Hydroporini we find that in Coelambus certain species show great differences in the
On Aquatic Carnivorous Coleoptera or Dytiscidce. 197
sculpture of the two sexes ; thus in Dytiscus parallelo^rammus (No. 416) the female differs from the male in that the punctuation is very much finer and less deep and the surface is dull, and on examination with a considerable power of the microscope it is seen that the dulhiess arises from the surface being covered with a very dense minute sculpture of a peculiar character, but a good deal similar to that which is seen on a larger scale in Meladema coriacea (No. 978). In another species of Coelambus (Dytiscus impressopunctatus No. 409) the phenomena are quite different ; here the female usually resembles the male in sculpture except that its punctuation is a little denser and finer : there occurs however very rarely a second form of the female (Dytiscus lineellus Gyll.) very different from the male, the surface being much more finely punctured and dull, and on examination with a considerable power of the microscope it is seen that the dullness arises from the surface being covered with an extremely fine obsolete sculpture of a reticulate nature such as is seen in Agabus. In the large genus Hydroporus the females frequently differ in sculpture from the males in a more or less conspicuous manner, and the difference is sometimes dimorphic {vide Hyphydrus memnonius No. 558). In the Agabini we find in the genus Agabus that sexual differences of sculpture are of frequent occurrence : the differences is sometimes slight as in the case of Dytiscus guttatus (No. 670) where the female has the sculpture a little coarser than in the male ; or it may be very considerable as in A. lecontei and A. griseipennis (No. 731 and 732), where the male is less but the female is more reticulate than is usual in the genus ; or the sexual differences of sculpture may be variable as in the cases of Dytiscus congener (No. 706), and Dytiscus bipustulatus (No. 751), and in these cases the variations seem to be, to some extent at any rate, dependent on the locality the specimens inhabit.
In the genus Copelatus a difference in the sculpture of the two sexes is very fre- quent, and the differences are in some species very great, in others only slight, and are quite as frequently found on the surface of the prothorax as on the wing-cases ; they usually consist of very short linear impressions, or fine, rather irregular scratches, dimorphism of the females is occasional, and when it exists is strongly marked, the female being either similar to the male in sculpture, or having a large area of the surface covered with an additional sculpture (vide Copelatus neglectus No. 841). It should be remembered that this genus is unique among the Dytiscidae by reason of the existence in it of a remarkable sculpture common to the two sexes : this non-sexual sculpture is of quite a similar nature to the sexual sculpture of other genera : the impressions on the thorax of Colymbetes sulcipeunis (No. 895) and the regular grooves on its elytra are just such as would (from what exists in the females of other genera) make one suppose them to be a female sexual sculpture, and yet they are equally developed in both male and female ; it is further interest- ing to note, that the females of the species I am alluding to (and of other Copelati)
have an additional sexual sculpture consisting of some extremely fine irregular
a D 3
198 On Aquatic Carnivorous Coleoptera or Dijtiscidce.
scratches on the wing-cases : it might therefore be roughly said that in Copelatus the males have gained a highly developed and pseudo sexual sculpture as ■well as the females, but that the latter have in addition to this a minute sculpture of a different character and confined to their sex. In the genus Coptotomus there exists a minute difference in the punctuation of a portion of the wing-cases, the punctures in the female being more elongate than in the male. And in Matus (Colymbetes bicarinatus No. 1)07) the females have the upper surface less shining than the males, owing to the existence of an extremely minute, almost impercep- tible, sculpture, which appears to be slightly more developed in the former of the two sexes. In the genus Rhantus sexual differences of sculpture are usually either absent or very slight, but in some species become important. Thus in Dytiscus bistriatus (No. 949) the female may on a very careful examination be seen to possess a slight develojjment of a more coarse reticulate sculpture than the male on certain spots situated along the lines of serial punctures ; and in Dytiscus pustulatus (No. 945) there are two small patches of a similar but deeper reticulation, one on the outer, and the second on the middle of the two rows of serial punctures, some little distance behind the base. In Pthantus anisonychus ? the sculpture on the outer and basal portion of the wing case, is a good deal deeper than in the male ; and in Rhantus discedens (No. 938) a deep coarse sculpture exists in the female on the part just mentioned ; while in Dytiscus notatus (No. 947) the whole of the wing- case, except a small portion at the outer and apical part, is covered Avith such sculpture : in this latter species however the female is sometimes smooth like the male, and individuals occur in which the sculpture is intermediate in extent between these two extremes. The peculiar sculpture of the genus Colymbetes exhibits some very interesting sexual peculiarities ; although in several species no appreciable difference between the sexes in this respect can be detected, yet in others very notable differences exist ; in Colymbetes sculptilis (No. 968) the peculiar transverse scratches of this genus are placed closer to one another in the female than they are in the male ; while in Dytiscus dolabratus (No. 971) and in D. striatus (No. 972) the female peculiarity is that the scratches are very much deeper than in the male ; in Colymbetes exaratus (No. 966) the strice or scratches are a little deeper and a little closer together in the female than they are in the male, and moreover in the former, the basal portion of the wing-case is rendered very dull, by the development of an excessively minute sculpture (consisting of very small longitudinal corrugations) which is quite absent in the other sex : in Colj'mbetes dahuricus* the sculpture of the female presents very exceptional peculiarities, it is decidedly finer than in the male, but it is more irregular, owing to the transverse scratches, being less elongate, and so running much more frequently into one another, and the surface is duller than in the male owing to the greater development of an exces-
* There is some little doubt whotlier the specimens of this species I here allude to are really the sexes of one and the same species.
On Aquatic Carnivorous Coleoptera or Dytiscidce. 199
sively minute interstitial sculpture, having a somewhat granular appearance. Passing to the Dytiscini we meet in that group with some most remarkable facts : in the genus Hyderodes the females are usually smooth and polished like the males, but they are dimorphic, inasmuch as a second form of the female is met with (apparently only rarely) in which the surface is excessively rough, the whole of the upper surface, except the head, being covered with deep coarse erosions or corru- gations, irregular in shape and direction. In the genus Dytiscus considerable dis- crepancies exist among the various species in the sexual sculpture ; in D. punctulatus the female has ten grooves on the basal portion of the wing-cases, and the whole of the rest of the upper surface, including the interstices of these grooves bears a close fine punctuation, while in the male the grooves are wanting, and the elytra are punctured only on the apical portions ; the female in this species has also even the undersurface rendered dull over a considerable portion of its area by the •existence of fine, short scratches or reticulations which are not found in the other sex ; in D. fasciventris the facts are similar except that the fine sculpture is less extensively developed, so that as regards this latter peculiarity the sexes are more alike than they are in D. punctulatus. In Dytiscus habilis, in D. hybridus, and in D. verticalis the females have no grooves on the wing-cases, but they differ from the males by a greater development of the punctuation on the apical portion of these parts, and also by possessing an additional fine punctuation on the lateral basal portion of the wing-case ; in D. verticalis this additional punctuation is very small and unimportant ; the females of these three species all possess too a con- spicuous special sculpture on each side of the prothorax ; in D. subllmbatus the female characters are approximately the same as in D. punctulatus and fasciventris, but in this species there also exist females differing in sculpture from the males only by their possessing a fine scanty punctuation on the lateral portions of the prothorax ; a similar condition is present in Dytiscus marginalis, the females differing greatly from the males by their grooved and much punctate surface, but individuals of their sex are found differhig from the males only by a slight punc- tuation on the thorax and a little greater development of that on the elytra. In D. circumcinctus we again find the females possessed of a grooved and much punctate surface, but here a second form of the female occurs quite without sexual punctuation or grooving of the wing-cases. The females in the genus Dytiscus differ then from the males by possession of a fine sexual sculpture, and by a groov- ing of the elytra, this latter feature is however not found in certain species, and in certain other species is sometimes present, sometimes absent, while the fine sexual sculpture of the females is more constant, but may also be occasionally quite absent, and is only present in its greatest development in such females as are sulcate. In the Hydaticides, we find certain species of Acilius possessing females with grooved wing-cases, but the grooves are very different from what exists in Dytiscus, they being but four (instead of nine or ten) in number, and furnished with a pubescence
200 On Aquatic Carnivorous Coleoptera or Dytiscidce.
which is confined to them ; and the spaces between the grooves are without punctuation, while in the males all the surface is punctate ; the females of the difterent species of this genus differ much in their characters from one another as to the development of the grooves and the punctuation and pubescence, and in Dytiscus sulcatus (of the genus Acilius) we find a patch of pubescence occupying a depression on each side of the prothoras.
In Thermonectes the surface is highly polished, and when the females possess a sexual sculpture, it consists of beautiful elongate punctures placed on the basal part of the wing-cases but not extending over a large part of their area. In Sandracottus the surface is very highly polished and the female is destitute of sexual sculpture. In Graphoderes the females are usually destitute of any sexual sculpture, except a slio-ht corrugation of the surface on each side of the prothorax, but in this genus we meet occasionally with a more extreme development of sexual sculpture than any found elsewhere in the Dytiscidee or indeed in the whole of the order Coleoptera, the surface of the wing-cases being rendered rough by a very coarse sculpture almost like tubercles (but not very different from what exists in Hyderodes) while the prothorax is covered with beautiful deep corrugations ; these exceptionally sculptured females are very rare and it has been thought they were only a second form of that sex in a species (D. zonatus) having usually smooth females, but I think more probably they belong to one or two distinct species of the genus.
The Hydaticini are insects with a very smooth surface, and many species are
without sexual sculpture, but others exhibit such well marked ; in this group the
chief seat of the sculpture is a portion of the area of each side of the thorax, but
sometimes also it is situate on the basal portion of the wing-cases ; it consists usually
of coarse, short, irregular impressions, but little connected with one another, and
sometimes there exists on certain individuals {vide in Dytiscus goryi No. 1020)
only one or two such impressions ; in a few New World species however the sexual
sculpture consists of a circumscribed, and remarkably well defined patch of quite
fine sculpture on each side of the thorax, formed by very closely placed fine rugae
{vide H. subfasciatus No. 1019). in the Cybistrini there is frequently present in
the female, a highly developed sexual sculpture, consisting of fine anastomosing
scratches, which frequently nearly cover the entire upper surface of the insect, but
in other cases are restricted to a much smaller area, sometimes this sculpture is
more developed on the thorax than it is on the elytra, while at other times we find
the reverse of this ; in many species of Cybister there frequently exists a sexual
sculpture so fine and slight that it can only be detected by a careful examination,
and many females are quite smooth ; in this genus, great variation of the sexual
sculpture is quite common in certain species : some species of Megadytes {vide M.
steinheili No. 1108) show a most beautiful sexual sculpture, the prothorax being
quite smooth, in great contrast to the wing-cases which are covered, at any rate on
On Aquatic Carnivorous Coleoptera or Dytiscidce. 201
their basal portion, with regularly placed elongate Impressions ; the largest of the Dytiscidce, viz., Cybister giganteus No. 1117 and Megadytes ducalis No. 1118, have no sexual sculpture whatever : in Homoeodytes we find the females of two species to possess a most excessively fine sexual sculpture, consisting of extremely delicate, short scratches, giving rise to a silky appearance on the basal part of the elytra.
From this imperfect review of the sexual sculpture of the Dytiscidas it will be gathered that much variety exists as to its character, and as to its degree of development, and that the occurrence of two forms of the female of a single species is not unfrequent in various portions of the family : there is considerable reason to suppose that the development of the sexual sculpture is to some extent connected with local and climatic conditions, and it may prove to be that it is of more frequent occurrence in temperate and cold climates than in warmer ones : if it serves any, or what purpose is still undetermined, but it is certainly amongst the most interesting peculiarities of the Dytiscidse, and of considerable importance in respect to its bearings on the questions of sexual variation and selection.
The peculiar sculpture of the species of Copelatus is of much interest and well deserving investigation. In some of the higher forms of the genus (Col. sulcipennis. Lap. e.(j.) it appears as very deep and regular stria3 or grooves, to the number of 10, 11, or 12 on each wing-case, and extending neai'ly their whole length, but becoming finer at the extremity, where usually some of them are more abbreviated than others. In other species no trace of such striation can be detected ; but in some of these cases the fine punctuation of the wing-cases assumes the form of very short fine scratches, which while they may be very distinct in the female, can sometimes scarcely be traced in the male (C. simplex, Clk.) The scratches displayed in such cases frequently assume a different direction on the hinder portion of the wing-case, becoming transverse there, when they have a longitudinal direction in front (Celina australis No. 806) ; while in some other species only the transverse posterior scratches exist (as in C ferrugineus).
In some of the cases where the elytra are truly striate, i.e. display elongate regular straight lines, it can be perceived that these are developed along the series of impressed punctures which are so constant a feature of the Dytiscidae ; this is not the case in those species, where the scratches are short, for they are not then naore developed near the series of the punctures than they are elsewhere ; so it would seem there are as regards the sculpture three ditFerent series of species, viz., 1, punctuation ordinary ; 2, punctuation elongate, and diffuse ; 3, punctuation -elongate, but only along the lines of serial (and secondarily of the interserial) punctures ; the punctuation of this latter category becoming in the most differentiated cases, highly developed symmetrical strise.
It seems impossible to believe that the development of this beautiful sculpture can have been determined by the action of natural selection ; preserving those individuals of a species in which it was more developed than in others, for iu
202 On Aquatic Carnivorous Coleoptera or Dytiscidoe.
certain species it is so excessively slight as to be almost iuappreciable, and the amount of variation to be detected in such cases being likewise very slight, one cannot believe in these slight differences of development exercising any appreciable effect on the existence of the creature. The sculpture however is most certainly cor- relative with sex even in its rudimentary forms; thus in Dytiscus agilis (No. 825) the female has the punctuation on the basal portion of the wing-cases more elongate than in the male ; and in the allied Copelatus atriceps the difference is exhibited in a still more rudimentary form : while in the very interesting C. dimorphus (No. 827) we have a species displaying true striation in a rudimentary form, and in the male in a much more rudimentary manner than in the female. And throughout the fjenus, wherever there is a difference it is that the females have the sculpture more developed than the males. We are entitled to believe then that whatever the influences may be that have brought about in Copelatus this peculiar sculpture, they are influences which have acted at first more strongly on the female than on the male, but that continued during a long period such disparity has disappeared, or tends to disappear.
It may be thought that these strice were of assistance in giving holding to the claws of the male, and that their direction enabled the male tarsi to arrive at the position most convenient for supporting the insect during the process of fertilization, and that the sculpture first gained by the female, was transmitted by heredity to the other sex. But these suppositions do not seem to me very satisfactory. One does not see why the females should be in the scratched species more different from the males, than they are in the striate ones ; for certainly the influence of heredity or sexual transference must be acting constantly, and not confined to the higher forms of development of the sculpture ; and it is excessively doubtful whether the highly developed sculpture is as useful for this sexual function as the rudimentary form : for the beautifully perfect straight stria; and grooves do not seem to me adapted at all for serving a useful purpose of the kind above suggested; the twenty-four deep broad parallel striae on the wing-cases of Col. sulcipennis can scarcely serve as agents to direct the claws to their requisite position, and I am at a loss to see in what other way they would be useful. Besides this, it is worthy of remark that in numerous species where the striation has attained a great development (similar in each sex) there exists in addition a very fine true sexual sculpture, peculiar to the female. The generalization of the facts in accordance with any theory of common origin, or of natural selection does not seem therefore to be warranted ; but on the other hand the presumption that each highly developed sjDecies has reached its develojjment by passing through a series of lower stages similar to such as are found still existing as the maximum of development in other species, is perfectly satisfactory and indeed iserristible to the imagination ; and the conclusion I come to on this subject of striation is that if we could gradually subtract from a higlily developed species its
On Aquatic Carnivorous Coleoptera, or Dytiscidw. 203
later acquisitions, and so go back step by step till we reached the striatiou in its rudimentary form, we should find it similar to, but not identical with, that of other species ; which is equivalent to saying that each species has had a truly separate line of development.
As a striking instance of the probable truth of this statement, I would point to the New Caledonian Copelatus aubei — the species in which the striation of the elytra reaches its maximum. The locality where it is found is remarkably rich in possessing very diffei'ent forms of the genus ; one of these forms C. interruptus (No. 847) has a striation of the elytra quite peculiar to itself; the striae are twelve in number but they are fragmentary, and slightly irregular in a peculiar manner. Now the only other species having so many as twelve strios is the C aubei above alluded to, and on examining this species one is brought to remark that on the portion of the wing-case where the stritie are usually least perfect — the apical portion — they show a fragmentary condition similar to what exists over the whole elytra in C. interruptus. In addition to this there is a true highly developed submarginal stria in C. aubei, and in C. interruptus, this stria is not present but is represented by a regular series of punctures, which only require extension to form the stria as seen in C. aubei. Now both these species are highly developed ones, found in one (isolated) locality, and the conclusion that their similarities are due to similarity of environment, and their differences due to a different condition of more original punctuation, is perfectly satisfactory to my mind. 1 conclude that C. interrujjtus and C. aubei have been developed from similar (but not identical) primitive conditions ; and that the serial punctures which seem to determine the lines of development of the striation, were less regular in the primitive ancestors of C. interruptus, than they were in the primitive ancestors of C. aubei, and that as the resultant of this the former species as we now see it has less regular striation than the latter has.
The Head in the Dytiscidee is of remarkably short, broad form, and is inserted on the prothorax in such a manner as to completely fill up the front part of the latter and so avoid any discontinuity of outline at the junction of the two ; it is considerably broader than long, and in Laccophilus, where the abbreviation is greatest, the width is about double the length ; its upper surface shows but little convexity, and is marked on each side, near the front, at a little distance from the inner margin of the eye, with an irregular depression or fovea ; and close to the suture with the labrum there is a transverse depression or short impressed line on each side; the clypeus is nearly always so completely joined to the posterior portion of the head, that the suture between the two is obliterated, although frequently its commencement can be traced near the eye, on each side, whence it extends inwards towards the anterior part of the irregular depression before alluded to ; in the genus Dytiscus (as also in Pelobius) the clypeal suture is distinct across all the width of the head, and in Meladema (especially in
TRANS. BOY. DUB. SOC , N S., VOL. II. ' ^
204 On Aquatic Carnivorous Coleoptera or DytisGid<B.
Dytiscus lanio) it cau also be traced nearly across the head ; on the other hand in Noterini, Laccophilini, and the Hydi'oporides, the clypeal suture is totally obliterated : in the genus Dytiscus where this suture, as already remarked, is unusually distinct it is in some species greatly deeper in one sex than in the other {vide Dytiscus dauricus. No. 998). The clypeus itself terminates over the labrum as a thin edge, so that the labrum continues the plane of the surface of the clypeus ; but in Hyphydrus we find an exception to this, the anterior part of the clypeus being deflexed at right angles, so that the labrum is placed on a different plane to that of the front of the head ; in this case a slightly raised line passes in a cui-ve from eye to eye, and marks oft' the small deflexed anterior portion of the clypeus from the rest ot the surface ; there are other members of the Hydroporides in which the deflexed anterior portion of the clypeus exists in a less distinct manner, {vide Dytiscus insequalis and allies in the genus Coelambus), and this is also the case with the raised margin, which exists in various degrees of obliteration ; or partial development, the middle part of this line is frequently absent, even when the lateral part remains distinct on each side, {vide species of Herophydrus and Hydrovatus) ; I have spoken of this margin as existing in various degrees of obliteration, but it is doubtful whether this is really the case, and perhaps it may rather be that it is in some cases partially developed ; this question can scarcely be determined without the aid of embryological research, but as the deflexed and margined clypeus is accompanied by a change in position of the parts of the mouth, inasmuch as they are in such case placed more on the undersurface, and as it seems to be an advantage to these predaceous beetles to have the parts of the mouth, more especially the mandibles and maxillse, quite in the anterior part of the head, I think the process of evolution is to bring about the diminution of the deflexed anterior portion of the head, and so permit the labrum, and the mandibles and maxillEB which this covers to be brought quite to the front of the head ; but from other considerations, which I need not here enter into, the reverse of this might be argued.
The depressed line, which I have described as existing on each side of the head close to the front angle of the clypeus, exhibits much variation, and is in the Hydro- porini frequently so indistinct as to be nearly absent ; in other cases these lines assume a greater extension so as to nearly join in the middle, and form a nearly con- tinuous line parallel with the front of the clj'peus (Dytiscus maculosus, No. 92, e.y.)-, in those species where the front of the clypeus is deflexed, there may be detected at its very front edge a transverse depression, which seems to have arisen from the extension of these Unes {vide Hyphydrus major) ; in the Noterides these clypeal depressions are usually placed nearer together, and so farther from the angles of the clypeus, while in Colpius they appear only to be represented by a very vague impression at a rather greater distance from the front edge. In the larger Dytiscida;, these depressions assume more the form of foveas than of lines, and are of remarkably
On Aquatic Carnivorous Coleoptera or Dytiscidce. 205
large size in Meladema. In the Agahini and some of the neighbouring genera, the clypeal depressions appear to be entirely wanting, but on a careful examination of Dytiscus bipustulatus it will be seen that the front of the clypeus appears to be provided with a fine margin, and on looking at other species of the genus (Agabus) it will be seen that the depression causing the margin arises from the extension in the transverse direction of the clypeal depressions.
The clypeus or epistoma is usually of a paler colour than the part of the head behind it, this is displayed in a very marked manner by the species of the genus Dytiscus, where the clypeus is yellow, and the front of the epicranium nearly black ; the darker colour of the epicranium frequently extends more or less on to the clypeus, so as to leave the anterior portion of this latter paler than the posterior part, but it is onl}' very rarely indeed that the anterior portion of the clypeus is quite black in colour.
The upper surface of the head shows no trace of any other suture besides the one existing between the epicranium and clypeus ; the transverse suture between the epicranium and protocranium which is very strongly marked in most of the Carabidae (in the natural condition concealed by the pronotum) being completely absent. The front of the epicranium on each side bears a well-marked irregular depression or fovea, which is occasionally more or less distinctly divided into two depressions placed one before the other (Meladema) ; these frontal fovese are entirely absent in the Noterides, and are only very indefinitely present in Hyphydrus, and some other Hydroporides, and are very much effaced in many members of the familv having a very smooth and polished surface, such as the Hydaticides and Laccophilini ; these fovese are more punctate than the rest of the upper surface, and carry some very fine and short depressed hairs. The large eyes encroach on the upper surface of the epicranium, and their inner edge is usuall}"- limited by a more or less punctate depression. The vertex, or portion of the epicranium behind the eyes is much broader than the front, and is covered at the sides by the angles of the prothorax ; thus, the eyes notwithstanding their large size are not prominent, and the breadth of the head behind the eyes is as great (or very nearly as great) as it is across the eyes, this being contrary to what exists in the Carabidte, where the greatest breadth of the upper surface lies on a line drawn between the convex portions of the two eyes ; there is no trace of any constriction behind the eyes. In the Colymbetides the surface of the head, when dark in colour, is usually marked by two more or less definite paler spots placed between the eyes, these frequently become united into one, and in Dytiscus are not only united, but placed so as to form an angular mark on the middle of the head ; and in the Hydaticides, the head is usually pale but with dark vertex, and angular dark marks in front, of variable extension according to the species examined.
The antenna is inserted more or less conspicuously on the undersurface of the head ; the anterior angle of the epicranium being inflexed, the cotyloid cavity for
2 E 2
20G On Aquatic Carnivorous Coleoptcra or Dytiscidce.
the anteuna is situated on it, and in front of the cavity the side of the clypeus i^ likewise a little infolded ; in the case of Eretes (and in a less striking manner in the other Hydaticides) where the eyes are lars-e, and the anterior parts of the head are very much reduced in size, the antennte are inserted rather on the upper than the under surface, the antennal cavity being quite visible from the front and above, owing to the diminution in size and the comparatively slight folding of the front angle of the epicranium ; on the other hand where the front of the clypeus is deflexed or inflexed, as in Hyphydrus, Qiieda, Hydrovatus, Pachydrus, the antennal cavity is situated quite on the undersurface of the head, the point of insertion being invisible when the insect is looked at from the front and above. The surface of the eye likewise is curved downwards, and continued on to the undersurface of the head so as to form a considerable portion of its area; between its inner termination and the elongate maxillary cleft of the buccal cavity there intervenes a naiTow space, with a very polished surface, which serves as a groove for the lodgment of the base of each antenna when these organs are placed in repose on the under surface of the head. The gula (piece basilaire of the French authors) is broad and distinct, each of the sutures separating its sides from the epicranium, terminates as in the Carabidte, in front by a deep impressed puncture ; this puncture is placed at a considerable distance behind the mental suture, and from it there is directed outwards a fine but distinct suture, extending as far as the maxillary cleft ; in this manner a large transverse piece is situated behind the meutum, but this jnece in the middle is not separated from the gula by any suture, and it is doubtful whether it should be considered a lateral exjiansion of the front of the gula (which in such case is T shaped) or as a submentum (piece prebasilaire of the French authors) : this same formation exists in the Carabidse, but is there excessively obscure owing to its being of smaller size, and limited by more or less obsolete sutures.*
Behind the eye there is a transverse suture or mark, which is clearly the remnant of the suture seen in the Carabidte as separating the epicranium from the protocranium : in a few cases this suture is very obsolete, thus in Dytiscus duodecimpustulatus (No. 462) it can scarcely be traced : it occupies much the same position in relation to the hind mai'gin of the head as it does in Carabidfe, but is much nearer to the anterior part of the head than it is in that family, though it is more than probable that there exists much difference amongst the members of the Carabidse in this respect ; indeed T find that in Carabus violacens, the suture in question is much nearer to the front that it is in Harpalus caliginosus. The relations of the parts near this suture may be summed up by saying that in the Dytiscidse as well as in the Carabidse the cheek is separated from the temple by a transverse suture, but that the cheek is very much shorter in the Dytiscidae than
• I believe myself that this transverse front bar of the gula is quite liomologous with the submentum (piece prebasilaire) of some other Coleoptera.
Oil Aquatic Carnivorous Coleoptera or Dytiscklce. 207
it is in the Carabiclte, although the temples remain of approximately similar mao-ni tudes in the two families.
There are but little variations in the structure of the head within the limits of the Dytiscidee ; in Amphizoa the head is not so short in proportion as it is in the other members of the family, and the antennal cavity is not quite so concealed, and the transverse sutures of the under surface are extremely obsolete ; these differences are however but slight, and in other respects the head of Amphizoa is essentially similar to that of the Dytiscidte.
The most important characteristics of the Dytiscid head besides its short, broad form, are first, its great extension transversely behind the eyes, so that its shape is far fi'om being cylindrical, and second, the great extension of the eyes on the under surface, so that they approach very near to the maxillary cleft : to these may be added as subsidiary, the broad gula, and the existence of the protocranial suture on the under but not on the upper surface.
In the genus Pelobius we meet however with insects possessing a head dlft'erent from that above described as characteristic of the Dytiscidas ; its posterior portion is cylindrical and not laterally dilated behind the eyes, which are prominent and convex ; there is no jirotoci'anial suture either on the upper or under surface, and the anterior angles of the clypeus are more produced or acute than in any other of the Dytiscidfe : on the other hand the gula is broad, and the eyes extend on the under surface near to the maxillary cleft, these being points characteristic of the Dytiscidse. Some of the members of the Pseudomorphini, one of the aberrant groups of the Carabidfe— ajiproach the Dytiscidse in several details of the structure of the head ; the antennse in Silphomorpha — one of the genera of Pseudomorphini — are inserted far on the under surface of the head ; the eyes are not prominent and approach on tlie under surface near to the maxillary cleft ; and the protocranial suture is distinct on the under surface but not on the upper surface ; while the general form of the head is broad and flat, and departs very far from the cylindrical form ; it is how- ever comparatively narrow behind, the cheeks are very much longer than the temples, and there is a very deep and remarkable antennary groove between the eye and the maxillary cleft : this head, therefore, though far from agreeing with that of the Dytiscidse, has the insertion of the antennse, and the development of the antennary groove an exaggeration of what exists in that family, and is thus in these points more different from the normal Carabidre than are the Dytiscidse. The Haliplides have a head which is remarkably deep in the vertical direction, the insertion of the antennse is conspicuously exposed on the upper surface of the head, there being no trace of that folding under of the antero-lateral portion of the epicranium so universal in the Dytiscidse (but which however is wanting in Eretes); the protocranial suture is distinct on the under, wanting on the upper surface, the cheek is very short, the temple large, and the gula broad, characters of the Dytiscidse : and the submental piece (piece pi-ebasilaire) is of large size, although its
"208 On Aquatic Carnivorous Coleoptera or Dytiscidce.
sutures are obliterated, so that its extent can only be determined by the two large punctures which mark the anterior limits of the lateral gular sutures.
The eyes in the Dytiscidee are always very large, and are placed at the sides of the head, but also encroach largely on its upper and under surfaces ; they are verv smooth and very finely facetted, but in this latter respect there exist some slight differences, as may be seen by a comparison of Agabus and Hydrovatus, with Eretes and Laccophilus, the facets being distinctly larger in the former genera. When looked at from the front of the head the curved outline of the eye is frequently much indented or emarginate just behind the antenna, by the side of the front part of the epicranium ; but in other cases the outline is preserved nearly intact at this point, so that the eye may be described os possessing in front a circular inner margin : Agabus and Hydrovatus may be referred to as exhibiting the former, Eretes and Cybister as exhibiting the latter structure ; as might be expected inter- mediate forms occur as regards this point, such are seen in Laccophilus, and even in Dytiscus.
The labrum is usually quite conspicuous, and placed on the front of the epistome or clypeus in such a manner as to continue its curve ; but in a few p-enera it is more or less withdrawn to-the under surface of the head, and is corres-
O
pondingly inconspicuous ; in Queda it is comjaletely concealed, and in Pachydrus, and numerous species of Hydrovatus its front margin alone can be seen ; in Hyphydrus although visible and exserted, it is placed more on the under surface of the head ; this position of the labrum depends entirely on the formation of the front of the epistome, for the labrum being always attached to the front edge of this part, varies in position according as this is inflexed or not ; thus in Queda and Pachydrus the front of the epistome is quite doubled or folded under, and thus carries the labrum back to the under surface of the head, while in Hyphydrus the front of the epistome, though not completely doubled under, is placed at right angles to its posterior part, and the labrum assumes a corresponding position ; Coelambus is a variable genus as regards the form of the front of the clypeus, and the ex- posure of the labrum. In its form the labrum shows but little variation; it is always strongly transverse, greatly broader than long, and its front edge is more or less emarginate in the middle, the emargination varying much as regards its breadth and depth ; the curve of the emargination is fringed with cilite implanted in a transverse groove: in Queda where the labrum is concealed, the whole of its anterior edge appears to be set with elongate fine cilioe, and in Hyphydrus the cilire extend nearly across the w^hole breath of the labrum, but in the great majority of the family the rule is that the emargination of the labrum is confined to a comparatively small width in the middle, and the cilife are quite short. In the Noterides the emargination is very slight, and the cilise are quite wanting : in Dytiscus latissimus the emargination is very slight. The labrum is always yellow in colour, and its upper surface is completely destitute of the isolated exserted setse which are so con-
On Aquatic Carnivorous Coleoptera or Dijlisciche. 203
spicuous in the Carabid^e andCicindelidee. InPelobius theemarginationof the labrum is destitute of the cilise found in most other Dytiscidfe. In Amphizoa the middle of the labrum appears from above ratherproniinent than emarginate, butwhenlooked at from in front and beneath, it appears very much the same as in the ordinary Dytiscidfe, but is destitute of the central series of ciliaj. On the inner face of the labrum in the middle in front there is in the Dytiscidee a prominent tiuangular space, densely covered with pubescence, and this exists equally in Amphizoa, but is absent in Pelobius and the Noterides, while in the Carabidse the rule appears to be that there is a glabrous space on the middle of the labrum, with pubescence on either side, completely the reverse of what prevails in the Dytiscidae.
The mandibles are always short, and in the condition of repose nearly or quite concealed by the short labrum, they are very broad, their outer edge forms a verv regular convex curve, while the upper and inner face is much twisted and is irregular in form, and the apex is turned into a broad somewhat sharp edge, which is a little prominent at each of its two corners ; the under surface is flat and bears along ils inner edge a band of fine, dense, very short pubescence, and there is also usually an isolated patch of similar pubescence near the apex.
The maxillaj are provided with two lobes, the outer one of which is quite slender, and is divided a little below the middle so as to form a two-jointed palpus ; the terminal joint is longer than the basal one, and is slender and cylindrical, usually a little acuminate towards the extremity, and the division between it and the basal joint is very perfect, so that the outer lobe of the maxilla is a perfect two-jointed palpus ; in Pelobius however the articulation between the two joints is imperfect, and the terminal joint instead of being cylindrical is broad and flat and has a large angular prominence on its inner edge, and an acute apex, so that the transformation of the maxillary lobe into a jointed palpus is very much less complete in this insect than it is in the ordinary Dytiscidse ; in Amphizoa the outer maxillary lobe is curved and slender, but it is flattened and not cylindric, and the only trace of a division is the existence of an emargination or notch on the inner edge, so that the condition of the outer maxillary lobe in this insect is very much that of the Gyrinidaj. The inner lobe of the maxilla is always strong and horny, and has an acute, bent inward termination, which plays the part of a hook in holding the food ; this terminal portion is always free from ciUse, but the lower portion of the lobe, along its inner edge is always fringed with ciliae. These ciliaB are subject to much difference in the family ; in Noterus there are only a few of them placed at a distance from one another, while in other cases {vide Eretes) they are dense and elongate ; on the upper face of the maxilla it may be seen that the inner lobe has at its base internally a membranous or semicorneous space, and it is usually on this space that are placed the longest and finest ciliae ; it is rare that the cilise of the maxilla are very fine, but they are frequently very coarse, as may be seen in Dytiscus, where some of the terminal ciliae are so stout as to form really powerful
210 On Aquatic Carnivorous Coleoptera or Dytiscidce.
hooks. In Pelobius the cilise are nearly entirely absent, the mner edge of the^ maxillary lobe bearing merely two short, fine sette ; in Amphizoa the ciliaj are abundant, and the lower portion of the lobe is furnished Avith dense fine pubescence, while the upper portion bears stouter cilite, a considerable space nearly free from cilise intervening between these two grovips.
The maxillary palpi are four jointed, the basal joint shorter than any of the others but quite distinct ; the second and third joints arc nearly similar to one another in length ; frequently the third is a little longer than the second ; the fourth joint is a little, or considerably longer than either of the others ; occasionally the terminal joint is notched or emarginate at the extremity (Noterides). The uiaxillte and their palpi are entirely destitute of any pubescence or seta3 (with the exception of the ciliai of the inner lobe). The maxillaj of the Dytiscidte are similar to those of the Carabidae, except in the absence of any outstanding setse and in the less elongate second joint of the palpi ; in this latter respect the Pseudomorphini agree with the Dytiscid*.
The mentum is a transverse corneous piece, of very hard consistency, articulated by a distinct suture vnih. the piece behind it ; this suture is slighty curved or bisinuate, or nearly straight, but is quite obsolete in Amphizoa : the sides of the mentum are rounded in such a manner that the greatest width is at the base ; in front there is a large piece cut out of the middle for the reception of the ligula, so that the mentum is thus made to consist of two lateral lobes, projecting from a transverse basal piece ; the base of the excision is not straight, but projects farthest forwards in the middle, owing to the excision being deeper on each side in adaption to the piece supporting the labial palpus ; there is thus formed a slightly projecting, more or Itss distinct central lobe at the base of the excision ; this lobe is variable in form, and is usually short in proportion to its width, and with a rounded front edge ; but is sometimes emarginate in front ; when the labial palpi are approximate to one another this lobe is narrow ; or nearly absent {vide Deronectes) and may form a small acute tooth ; such a tooth exists in Noterus but has in that genus a bifid extremity. The side wings or lateral lobes of the mentum vary a little but not much in form ; in Hyphydrus their front termination forms a more acute and elongate angle than it does in other groups. In Pelobius the mentum is very much shorter than in any other Dytiscidee, so that the very broad excision is very short, and the very broad lateral lobes are extremely abbreviate ; in Amphizoa the mentum is remarkable for its large size, the basal transverse piece being very elongate, but in other respects it is formed as in the ordinary Dytiscida^.
The ligula is large and very conspicuous, it appears externally as a quadrate or transverse horny plate, filling up and projecting beyond the excision of the mentum : its free front margin varies a little in form, it may be nearly straight as in Cybister, or slightly rounded or even very feebly prominent in the middle, so as to be bi-emarginate, and is usually furnished \\ith elongate very fine ciliie, placed some-
On Aquatic Carnivorous Coleoptera or DytiscidcB. 211
times so as to form two or more fine pencils, but sometimes extending along the whole margin in an equal manner : the outer surface of the ligula is impressed on each side for the accommodation of the labial palpi, so that along the middle it is more or less bulged or prominent : on the inner face the ligula is membranous, except that there is a transverse horny strip at the extremity. The paraglossse are closely united to the ligula, and do not project beyond it ; on the outer face of the ligula they are difiicult to detect, but are nevertheless visible on each side, as forming a more transparent corneous outer edge to the ligula ; but on the inner face of the ligula they are very conspicuous, as each of them has its inner margin, fringed with an elongate series of very fine, dense cilise. The supports of the labial palpi are conspicuous and at their base are attached to the outer face of the ligula, and covered by the front portion of the central piece of the mentum ; their extremity is on the contrary ver}^ conspicuous, and appears in the form as it were of a stout basal joint to each palpus ; this joint bears a deep notch on its antero-external face and thus the palpus is capable of being bent back over the mentum as well as moved in a forward and dependent manner. The labial palpi are of moderate length and are three- jointed : the basal joint is considerably shorter than either of the others, while the second and third differ but little from one another in length (except in the Noterides) ; the second joint frequently bears a minute prominence on the middle of its inner face causing it to appear faintly bi-emarginate on this fiice : the apical joint is usually simple and cylindrical, but may be dilated, and bear a notch (Coptotomus).
In the Noterides the labial palpi are peculiar, the basal joints being short, while the apical one is large, more or less dilated, and bears a notch : in Hydrocanthus this dilatation of the terminal joint of the labial palpus becomes extreme.
The labial palpi and the outer face of the ligula, are, like the maxillary palpi, polished and shining and destitute of any exserted setae. In Pelobius the ligula shows much diff"erence from that of the Dytiscidte, it is short, and its anterior portion is flexed upwards so as to assume an obliquely vertical direction, this semi-vertical portion is rather coarsely punctate. I can detect no paraglossse, and the supports of the labial palpi are not emarginate. In Amphizoa the ligula does not project beyond the lateral wings of the mentum, owing to the great development of these latter, but its structure is quite that of the Dytiscidse.
The parts of the mouth in the Dy tiscidfB offer two points of interest — first, the great uniformity they show throughout the family, and second, the way in which the different pieces are coadapted so as to close together quite completely and exclude the entry of water : the mandibles are curved so that they close under the labrum, and of the maxillse only the very hard, hooked, extremities are at all exposed, the ligula and mentum completely closing the underside of the mouth ; notwithstanding this close packing together of the parts of the mouth the whole of the maxillary and labial palpi are entirely exposed — more completely, in fact, than is usual.
TBANS. ROT. DUB. SOC, N.S., VOL. II. 2 F
212 On Aquatic Carnivorous Coleoptera or Dytiscidce.
The small amount of variation exhibited by the trophi throughout the family is perhaps largely conuected with the different parts being thus closely packed together. However this may be, the absence of any conspicuous developments of the trophi is remarkable, especially when we recall the fact that these parts in the Dytiscidce are extremely similar, except in some minor details, to those of the Carabidoe, and remember the great modifications of these parts that the latter family displays; ifthelabrum of Carabus, where the ligula is merely a small semi-mem- branous piece connecting the anterior parts of the large paraglossfe, be contrasted with that of Authia, where the ligula is a very large, elongate, horny lobe, at the base of which on each side is attached the insignificant paraglossa, the great difference here displayed will render very striking the uniformity of the correspond- ing parts throughout the Dytiscidas.
The parts of the mouth show no important differences from those of the Carabidae, but there are two or three comparatively unimportant details as to which a suffi- ciently constant di;i'erence exists to make them worthy of mention ; these are — first, the complete absence of outstanding setae from the palpi or other parts of the mouth : second, the fact that the second joint of the maxillary palpi is only about equal in length to the third, whereas in the Carabidfe in the vast majority of cases the second joint is considerably longer than the third; and, third, the deep emargination of the extremity of the lower face of the supports of the labial palpi. These characters are not, it will be admitted, of great importance, and as regards the second of them it may be mentioned that, in the Pseudomorphini, the second joint of the maxillary palpus remains, as in the Dytiscidse, scarcely longer then the third joint. The undivided external maxillary lobe seen in Amphizoa is also paralleled among the Carabidse (Callistus, and one or two other genera).
The A^JTENN^ of the Dytiscidse are always eleven-jointed, and are usually slender with the joints elongate, the second being, however, frequently shorter than the others, and the joints from the third to the apex each a little shorter and more slender than its predecessor. Although this is the structure in the larger and higher forms, there is a considerable diversity in other parts of the family. The Noterides are especially peculiar in this respect, their antennse, being short and more or less distinctly incrassate in the middle ; in this group even when the antennae are apparently .slender and simple as in Hydrocanthus, it will be found that the middle joints (joints G to 8) are a little broader and longer than the following or than the preceding ones. In the Hydroporides the rule is that the second joint is as long as, or longer than the third, and the antennse are only as long as, or shorter than the head and thorax, and not remarkably slender, while in the Hydi'ovatini and Hyphidrini and Bidessini it is not unusual that they should be feebly serrate internally. In the Colymbetides the antennae are slender and filiform or setaceous, but never become very remarkable on account of elongation and tenuity, and in the earlier genera of the group, are usually comparatively thick ; it is in the
On Aquatic Carnivorous Coleoptera or Dytiscida. 213
Hydaticides and Cybistrini and Laccophilus that the tenuity of the antennae becomes extreme ; in some of the larger species of Cybister, such as C. owas, the antenna are so long and slender, and the joints are so very feebly connected together that it is difficult to find in collections a specimen with these organs unbroken. In some genera of the family the males have the antennas distinctly different from the female, the difference being usually that the middle joints are more or less dilated and assume peculiar shapes; this sexual difference is seen to its greatest extent in the genus Noterus ; in some Hydrovatini (especially in Hydrovatus aristidis) the antennas of the male are remarkable for their form; in the males of some species of Sternopriscus the form of the antennas is excessively bizarre, and two or three species of Agabus (Dytiscus serricornis No. 755, e.g) are remarkable inasmuch as the apical joints of the antennae are dilated in the male.
The most striking peculiarity of the antennae of the Dytiscida?, is that they are quite free from setae, from sensitive pubescence or from pvmctuation, their integument being quite shining and polished. In Pelobius however the very large basal joint of the antennae is very distinctly punctate ; and in Amphizoa the antennae are still more considerably punctate, the four or five basal joints showing an extensive though rather irregular and obsolete punctuation, and the following ones being each in succession more sparingly punctured, so that only the apical joints appear entirely glabrous and shining.
This condition of the antennae is one of the most important of the distinctions between the Dytiscidas and Carabidae, the members of this latter family having the antennae setose, and the joints, except the three or four basal ones, covered with an excessively fine, short, and dense pubescence which, in conjunction with numerous minute pits on the surface of the joints, makes them dull, and is considered to be an external apparatus of sensation. This glabrous condition of the antennae in the water-inhabiting Dytiscidae is therefore of interest as helping us to interpret the function of the antennas in insects — a very complicated and difficult subject.
There are certain facts which render it probable that the simple condition of the antennee of the Dytiscidre is not to be attributed directly to their aquatic existence, but is rather correlative with their not being in contact with the atmosphere. The species of Hydrophilus (aquatic beetles of the family Hydrophilidfe) breathe in a very peculiar manner, by protruding from the surface of the water the three apical joints of their aatennte, and by their means carry down a supply of air to the under surface of the body ; now it is a remarkable fact that these three apical joints thus exposed to atmospheric influences are covered with a very dense and fine sensitive pubsscence, while all the rest of the antenna which remains in the water is completely glabrous and shining like the antennae in Dytiscidae. There are, moreover, certain Carabidae in which the sensitive pubescence of the antennae is nearly absent, and it is interesting to note that there is reason to suppose that these are species which in the perfect state are nearly or completely withdrawn
214 On Aquatic Carnivorous Coleoptera or DytiscidcB.
from the air on account of their completely subterranean existence ; we may therefore conclude that it is under the influence of much exposure to the atmosphere that the sensitive pubescenca on the antennas of beetles has been developed. The more elongate exserted setae present in the Carabidre but absent in the Dytiscidaj have probably a totally different function from the fine sensitive pubescence ; it is quite possible that they are tactile organs similar to the whiskers of the cat, and that they are absent in the Dytiscidpe because the resistance of the water when the insect is active would move the setre, and, if they are as I have supposed sensitive to pressure, render them a nuisance to their possessor. The study of these structures falls however to the Carabophilist, but I must remark in connexion with the question of the relations between the Carabidse and Dytiscidse, that in the former family there exists considerable differences in the clothing of the antennie ; tlms in the rare and anomalous Trachypachys there is a complex and symmetrical system of antennal setaj, but sensitive pubescence seems to be entirely absent ; while in other cases, as in Anthia, there is a beautiful development of the sensitive pubescence and a nearly complete absence of setse. In the Scaritidae and Broscidiie the four basal joints of the antennae are glabrous (except that in some Broscidaj porosity begins at the apex of the fourth joint) but on the other joints the rule is that the porosity is confined to the edges of the more or less compressed joints, the flattened sides being glabrous. Exceptions occur in %vhich the porosity covers the whole of these joints, and there are other exceptions —these much more remarkable — in which the porosity is almost entirely absent. It is especially noteworthy that the forms presenting this latter peculiarity are those which from their form and appearance, are probably most completely subterranean in their habits, such as Monocentrum, Teratidium, Neocarenum, Passalidius, and Metaglymma. In Passalidius there are neither punctures nor hairs, but there are grooves on the edges of the joints. In Neocarenum there exists scarcely any pubescence but a few coarse punctures. In Scarites excavatus there are large punctures mixed with the fine porosity on the edges of the joints. Thus we see that in the subterranean Carabidae, where the antennre are less subject to atmos- pheric influences than the other members of the family, the antennce have their sensitive structures less largely developed than usual, and that in the most completely subterranean forms of the family the antennas approximate to those of the aquatic Dytiscidae, without however being so completely simple as in the water beetles.'^'"
* I am greatly indebted to Mr. H. W. Bates for furnishing to me the information above recorded about the sensitive structures of the antennae in the Scaiitini and Broscini. Mr. Bates has examined many species in the following genera with the results thus tabulated — Carcnum, Pasimachus Emydopterus, Euryscaphus, Carenidium, O.vylobus, Scaraphites, Crepidopterus ; in the above the compressed joints 4—11 are glabrous only along the middle of the flattened sides ; while in Monocentnim Teratidium, Neocarenum, Passalidius, Metaglymma, Brullea, the antenna; have a diminished band of jiorosity on the edges or are entirely smooth; in Scarites they are much as in Carenum but more, variable ; in Gnathoxys there are rather distant pores, and in Percosoma there is dense porosity.
On Aquatic Carnivorous Coleoptera or DytiscidoB. 2.15
As regards the slenderness of the anteniiEe in the Dytiscidse it may be remarked that the degree of tenuity may probably prove to be in direct ratio with the activity of the species : very slender antennoB are found only in the best swimmers, and we can readily understand that it is favourable to active locomotion that the antennce should be very slender and flexible, so as to stream back with facility along the under surface of the body during rapid motion, and thus offer no obstacle to progression.
The Pkothorax is at its base very intimately applied to the elytra and mesoster- num, and so accurately fitted therewith as to shut out the entry of water, although this great articulation still permits considerable movement of extension : in front it very closely and accurately clasj^s the head. It is always strongly transverse, its greatest length being, in consequence of the prolongation of the anterior angles, at the outside, but the breadth is usually twice, or more, its greatest length. The sides are generally gently curved, and the bi'eadth increases from the front angles to the base or very near it, the sides therefore diverging from the front to the base ; rarely this divergence is absent, as in Vatellini, less rarely the greatest width is across the middle, the sides being slightly contracted from thence towards the base (Amphizo;), Tyndallhydrus, Andex, many Deronectes, and a few Agabi). When the prothorax is broad at the base, the hind angles are well marked, frequently acute, and in Neptosternus they are produced so far backwards as to be spinose ; in other cases the hind angles are nearly rectangular ; in Deronectes frequently obtuse and rounded, and obtuse in a few Agabi : it is worthy of remark that the continuity of the outline of the thorax and elytra is very complete in the higher Dytiscidce, and in the exceptions where it is very incomplete it is apparently always or nearly always an accompaniment of an imperfect articulation of the prosternal process with the mesosternum and metasternum. Thus in Andex, Tyndallhydrus, and the Vatellini, where the pronotum is narrow, and not at all continuous in outline with the wing cases, the prosternal process fails to articulate with the metasternum at all, and is abbreviate in front of the middle coxge instead of being prolonged between them ; while in the genera of Hydroporini in which the outline of thorax and elytra are very discontinuous, the mesosternal foi'k is disconnected with the metasternum, and in those Agabini where the hind angles of the thorax are obtuse or rectangular, the articulation between the prosternal process and the metasternum is very imperfect (vide Agabus wasastjernte, A. cephalotes, &c.), only when the prosternal process is very perfectly held or fixed do we find the base of the thorax become very broad so as to completely continue the outline of the afterpart of the body, and render perfect the form for motion through the water : it is in the Hydaticides and Cybistrini that we meet with the most perfect outlines, and these are the groups where the prosternal process is most perfectly articulated with the metasternum. On the other hand in these most perfected forms the capability of extension or mobility of the prothorax from the afterbody is nearly completely lost.
216 On Aquatic Carnivorous Coleoptera or Dytiscidw.
The side of the pronotum is most frequently Hmited above by a raised margin ; this margin may be broad as in Hyderodes, or very fine as m Eretes, or may be entirely absent, as in the Cybistrini and numerous others, and all the grades of connection exist between these extreme instances. In a few cases the anterior edge of the pronotum is also finely margined as in Coptotomus and Lancetes, but this is rare, and a raised front margin is generally nearly absent, although usually more or less distinct traces of its existence may be seen in the part of the thorax near to the anterior angle. In Pelobius the anterior margin is conspicuously ciliate, and in Amphizoa it is furnished with a narrow strip of pale membrane ; in the other Dytiscidte it is destitute of either membrane or cilire, except that frequently the ciliae may still be seen on a small extent of the front margin close to the anterior angles, as may be well observed in Cybister ; these rudiments of cilite are usually confined to an extremely small space, and are very difficult to detect with- out dissecting the head from out of the thoracic cavity. The base of the pronotum in those genera having a visible scutellum is usually a little sinuate on each side near the outer angles ; in those genera where the scutellum is invisible the base is a little produced backwards in the middle so that a more or less distinct angle is formed. The surface of the pronotum never bears any setae, and has no distinct basal or other fovete, but very frequentl}^ a more or less distinct transverse series of punctures near the front margin but not reaching the sides may be seen, and also a second series near the hind margin, but more or less broadly interrupted in the middle : on the disc traces of a longitudinal channel may be seen in the shape of a fine short impression, and in Lancetes this channel becomes so elongate as to extend half ihe length of the thorax, while in Amphizoa it is entire : in Bidessus and in Sternoprlscus there exists a peculiar longitudinal plica or fold on each side at some considerable distance from the middle.
On \iewing the prothorax from the side it is seen that the sides of the pronotum form a more or less conspicuous roof-like expansion ; this expansion is a growth of the pronotum and does not include any portion of the prosternum, for on looking at the under surface it is seen that the junction between the pronotum and the pieces of the prosternum is not at the outside edge of the thorax, but entirely on its under surface, at a considerable distance from the edge. The prosternum proper is a trans- verse piece, connected with the episternum by a suture which is usually very distinct, but in Hydroporides is quite obsolete ; its lateral termination connects with the side of the pronotum by a very short suture which is frequently quite open in front ; although the sides of the prosternum proper are thus very short (in the longitudinal direction) the middle portion has a very large development, it takes a downwards and backwards direction between the anterior coxae and is prolonged behind them to articulate with the middle of the meso- and metasterna ; the prosternum proper may therefore be described as consisting of a middle longitudinal piece, and a lateral wing on each side of it in front. The longitudinal jiortion is of much impor-
On Aquatic Carnivorous CoJeoptera or Dytiscldie. 217
tance, and is more or less distinctly marked off into three jiarts, 1. The portion behind the coxse, this is called the prosternal process ; 2. The portion lying between the coxse, which may be called the intercoxal band ; and 3. The portion in front of the coxse, which may be called the anterior band. This anterior band forms on either side the front boundary of the coxal cavity, and at the outer side of the coxa is prolonged backwards, and articulates by aver}' distinct suture with the posterior side-piece or epimeron ; external to the coxa the anterior band is continued outwards, forming the side wing of the j^rosternum proper ; this side wing is always very short (in the longitudinal direction). The anterior band varies somewhat in length, according to whether the front coxje reach very near to the anterior edge of the prosternum or not ; it is extremely short in Hyphydrus but is considerably longer in most of the Noterides, and in the Colymbetides, Dytiscini, Cybistrini and Hyda- ticides it is of moderate length (except in Eretes where it is exceptionally short), and in these groups, which might be called the Macro-Dytiscidae, in consequence of their greater size contrasted with the other groups, it is thickened along the middle, though to a variable extent ; this thickening catises the under surface of the prothorax to be of a keel-like form, as may be well seen in the Cybistrini where it reaches its greatest degree of development ; in the Colymbetides this keel-like thickening is usually much less than it is in the Cybistrini and Hydaticides, and in the lower forms of the Agabini such as Agabus cephalotes (No. G63) and A. cordatus it may be seen in its rudimentary condition ; Coptotomus, however, is an exception in this respect to its allies. Passing now to the intercoxal portion of the prosteinum, we find that in the Hydroporides this part forms a slender band between the coxae, and that it attains its greatest elongation in Hyphydrus ; in ■some of the Bidessini (Pachydrus), a very peculiar condition of this baud exists, in front its margins are a little raised so that it is somewhat depressed or sulcata along the middle, but about halfway of its length it becomes suddenly very promi- nent or protuberant, and convex instead of sulcate, giving rise to the suggestion that the anterior and posterior portions are distinct pieces ; traces of this division of the intercoxal band into two portions exist in ntxmerous members of the Hydro- porides, inasmuch as the posterior portion is less flat transversely than the anterior, and that there is a more or less distinct tubercular prominence at the junction of the two parts {vide Dytiscus parallelogrammus. No. 416, and numerous other species of Coelambus and the allied genera). In the Macro-Dytiscidse the condition of the prosternum between the coxse is very different, for the thickening or incrass- ation of the middle which I have described when speaking of the anterior band, is continued between the coxse, so that here the prosternum projects beyond the level of the coxse, or appears as it were to be arched over them, not merely produced between them, as in the Hydroporides. Behind the coxse, the prosternum becomes broader, though sometimes only slightly so (Cybister, D}'tiscus, and many others), and is more or less prolonged backwards to form the prosternal
218 On Aquatic Carnivorous Coleoptera or Dytiscidce.
process ; this process differs greatly in its shape or form, and offers very important
and constant means of recognizing some of the genera and species : it is sometimes
very broad and shoi-t, the extreme of this condition being seen in Hydro vatus,
where the prosternal process projects but little behind the coxte, and is much
broader than long, and is moreover nearly truncate behind, the hind margin
showing however (when the prothorax is separated from the after-body) an obscure
angle in the middle ; the other extreme is seen in Laccophilus, where the prosternal
process, is very narrow, and is prolonged backwaids as a longer or shorter slender
spine ; a great number of other forms more or less intermediate exist between these
two extremes ; but in Neptosternus we find a very exceptional prosternal process,
the process itself being trisjMuose, and consisting of an elongate slender spine in
the middle, with a shortei', and very slender, divergent spine on each side. The
prosternal process is either flat in the transverse direction, as in most Hydroporides,
some Agabini (Platambus, &c.), or is more or less compressed (Ilybius), convex,
(Coptotomus) or even indistinctly carinate along the middle (Herophydrus) ; it is
margined at the sides, the margin being the backward prolongation of a similar
margin existing along the front of the coxal cavity {vide Cybister) ; this margin
sometimes extends for the whole length of the process {vide many Agabini, Dytiscus
fuscipennis No. 752 e.g.), or becomes slender and terminates before the extremity
(Dytiscus, Cybister and many others) ; sometimes the margin exists at the side of
the prosternal process, although it has become quite obsolete along the front of the
coxal cavity, and it appears to be the rule in th3 Hydroporides that this margin of
the coxal cavity and prosternal process is absent or only very partially developed :
the most remarkable condition of the margin of the prosternal process is that found
in Neptosternus, for on examining N. tridens in comparison with some other
Dytiscidse, it will readily be seen that the lateral spines of the prosternal process
are simply the lateral margins which are detached and divergent ; this peculiar
condition of the prosternal process in Neptosternus is well worthy of the attention
of entomologists who think that the doctrine of community of descent may be true
in the case of allied genera and species, if not in the case of all living beings ; for
we have here a perfectly isolated and most peculiar form of appendage, which there
seems every reason to believe must have been developed pari passu with the
prosternal process, for it is almost impossible to suppose that the margin was in the
ancestors part of the prosternal jarocess, and has since become gradually detached;
whereas we can well understand the prosternal process and its margins as being
each a distinct growth, amalgamated in other beetles, but distinct in Neptosternus ;
if such be the case how far back must we go in the ancestral record before we
could hope to find a common ancestor for the allied genera Neptosternus and
Laccophilus ? We must go back to the period when there existed no prosternal
process. But the existence of a prosternal process is an absolutely constant feature
of the Dytiscidse ; and must no doubt have been one of the earliest developed of
On Aquatic Carnivorovs Coleoptera or Dytiscidce, 21ft
the special characters of the family. So that the epoch to which we must remount before we can imagine a common ancestor for these two allied genera is enormously- remote : Avhile to believe that there may have been a common ancestor for the whole of the Dytiscidse becomes almost impossible.
The anterior coxal cavities are placed as a kind of excavation, one on each side of the middle piece of the prosternum, and vary in shape in confomiity with the coxoe ; they are imperfect inasmuch as they only very partially enclose the coxce, especially in those forms where the coxse are elongate as in Hyphydrus ; owing to their peculiar form and position, the cavities appear to be open behind, but this is not really the case, and in fact all the Dytiscidse have the coxal cavities closed behind ; the closing of the cavities is effected, as usual, by an expansion given off from either side of the posterior aspect or edge of the middle piece of the prosternum ; this expansion is in the Dytiscidfe nearly colouidess, and might be supposed to be membranous, or semimembranous in consistency, but it is really quite corneous : moreover it does not connect together the posterior extremities of the side pieces as is usual in the Coleoptera with closed coxal cavities, but is placed as it were somewhat in the interior of the prothorax, and so leaves the angles of the side pieces free and prominent ; this mechanism is highly important as it is, in conjunction with the articulation between the prosternal process and the afterbody, the means of securing a perfect articulation between the prothorax and the afterbody, and consequently of excluding water from the interior of the insect ; the prominent transverse portion of the bridge closing the coxal cavities, actually enters into the interior of the mesosternum (of which it has exactly the width) while on each side there is left a groove or depression into which the front edge of the mesosternum fits, the free edge of the prothoracic side piece playing over that of the mesothoracic piece. The coadaptation of the various parts of the posterior aspect of the prothorax, to corresponding parts of the afterbody and base of the wing-cases is extremely perfect and \exy complicated; proceeding from below upwards we have, first, the prosternal process stretching beyond the mesosternum to be received in a metasternal groove ; directly above the prosternal process we see a considerable protuberance or prominence which fits into the fork of the mesosternum ; then come the posterior aspects of the cox^ which fit into facetts on the face of the mesosternum, and on a still higher level we have the transverse bridge closinsf the coxal cavities which fits into the interior of the mesosternum, while on the upper surface we find that the base of the mesothorax and scutellum are shaped so as to allow the hind margin of the pronotum to overlap and accurately fit them, while the shoulders of the wing cases are prominent, and rest on an expansion of the posterior face of the pronotum which ^ beautifully sinuate and emarginate to facilitate the coadaptation. This jointing is so perfect in the higher forms such as Cybister, that if after the prothorax has been detached from
the afterbody an attempt be made to replace it in its natural position, this
a G
TEAKS. EOT. DUB. SOC. , N.S., YOL. II. ,
220 On Aquatic Carnivorous Coleoptera or DijtisciclfB.
s very easily effected ; and it will then be found that the thorax retains its position in spite of considerable efforts being made to dislodge it ; it may however be instantly separated by the point of a needle being thrust under the apex of the prosternal process. The most completely fixed prothorax is found in the Noterides, and more specially in the genus Hydrocanthus, where the excessively broad, truncate prosternal process is so accurately fitted to the metasternum, that the two parts look as if they were soldered together. The epimera and episterna of the prothorax sometimes exist in the Dytiscidae with their sutures distinct (as in Acilius and Cybister) so that the shapes and the comparative areas of the two pieces can be readily distinguished ; but in the Hydropo- rides the sutures are usually obsolete. When they can be distinguished it is seen that the larger part of the side of the prosternum is formed by the episternum, the epimeron being a comparatively linear piece : the front coxal cavity is formed how- ever by the junction of the epimeron wnth the anterior band of the prosternum proper; the epimeron sends off a prolongation for this purpose along the border of the coxal cavity, which entirely cuts off the episternum from forming any part of the articular cavity.
The structure of the prothorax in the Dytiscidse is essentially similar to that of the Carabidse, the most important differences being the great development and per- fection of the prosternal process, and the mode in which the coxal cavities are closed behind. There are some Carabidse which possess, however, a highly developed pro- sternal process like the Dytiscidae {vide especially Cyclosomus): and in Trachypachys and Systolosoma of the Carabidfe, the structure of the prothorax approaches very much to what exists in the Dytiscidse, and indeed shows comparatively little differ- ence from that of Amphizoa. These three genera, Trachypachys, Systolosoma and Amphizoa may truly be said to show a structure of the prothorax intermediate between that of the Dytiscidfe and Carabida?. The prothorax of Pelobius is very different — quite different one might say — from that of Amphizoa, Trachypachys and Systolosoma, and in certain respects is formed like that of the Carabida?, though on the whole it must be pronounced to be a very different prothorax from that of any of the Dytiscidse or Carabidse; its general form, apart from its acuminate and large and remarkably elongate prosternal process, is rather that of the Carabidse than of the Dytiscidse, and the very conspicuously ciliate front margins are found in no other Dytis- cidse but are the rule in Carabidce ; leaving the general shape and prosternal process out of consideration we find that it has the characters of the Dytiscid prothorax, but in a quite peculiar form; the middle band of the prosternum assumes immediately at its front margin a directly vertical direction, and then curves backwards arching over the coxse which are very elongate ; this marked prominence and incrassation of the middle of the prosternum is found only in some of the higher Dytiscidse, such as the Cybistrini ; the intercoxal band shows on each side a raised margin, which is quite independent of the usual margin of the coxal cavity and prosternal process, but ap- pears to be really similar to the fine raised margin found in a similar position in Cy-
On Aquatic Carnivorous Coleoptera or Dytiscidce. 221
bister scutellaris (No. 1 1 U 1) : tkese facts suggest that it is j ust possible that the remote and less perfected ancestors of Cybister scutellaris may have possessed a prothorax somewhat similar to that of Pelobius. The prothorax of the Haliplidte is likewise intermediate between the Carabidte and Dytiscidse, its general shape being Carabideous, while it possesses a large prosternal process and the coxal cavities are closed behind, by what may be called a horny membrane so as to be in these respects nearer to the Dytiscida^ ; it shows however characters extremely difierent from Pelobius, the front coxae being short and globular, instead of elongate as in Pelobius, the anterior margins destitute of ciliae, and the prosternal process broad, and quite truncate behind instead of acuminate.
In the Carabidte the peculiar genus Omophron has been supposed to be in respect of its prothorax an approximation to the Dytiscidce, but this is not really the case, for the coxal cavities are closed behind in a manner thoroughly Carabideous, and quite different fi'om what we find in Dytiscidas, Haliplidoe, Pelobius, and those Carabidaj which ajDproximate them (Trachypachys and Systolosoma). The prothorax of Silphomorpha (Cai-abidee) in some respects approaches that of the Dytiscidae, but in others it is very different ; the broad roof-like sides, and the existence of a prosternal process, and the absence of ciliation on the upper front margin are the points of approximation, while the form of the prosternal process, the mode of closure of the coxal cavities, and the very short coxae, are very different from what we find in the Dytiscidse. If, however, the prothorax of a Carabus, of an Agabus and of a Silphomorpha were taken for comparison, I think we might quite as justly conclude that the thorax of the water beetles was intermediate between those of the two Carabidte, as that any other linear arrange- ment of the three was correct.
In concluding this slight sketch of the structure of the prothorax in the Dytiscidse and some of the forms that approximate to the family, I may be pardoned for pointing out that any attempt to account for the resemblances we find to exist by a theory of heredity or community of descent only leads one into a maze of jDerplexities, which becomes the more distressing the farther we pursue it. If, however, we attempt to account for the resemblances as being adaptations, either by evolution or special creation, to the conditions of existence, or the environment, we find them qmte comprehensible ; in the beetles inhabiting water, viz., Dytiscidse, Amphizoa and Pelobius, it has been necessary to exclude the water from the interior of the body, and to fix the prothorax solidly so that it shall not be moved from its place during locomotion ; tlms we find perfect coadaptation of the parts to suit those of the afterbody, and the locking of the parts together by a prosternal process reaching to a part behind the articulation required to be rendered secure ; moreover in the higher forms which move much and rapidly in the water we find all projections and prominences that would impede forward motion diminished or removed, and it js possibly for this reason that there is an absence of the ciliae of the anterior margin.
2 G i
222 On Aquatic Carnivorous CoJeoptera or Dytiscidce.
Thus in the very different Pelobius, Amphizoa, Dytiscidse, and Haliplidoe, we find a similarity of structure of the prothorax existing, not because of community of descent, but because water ha\dng to be excluded from the interior of the body, the simplest and best method of accomijlishing that end has been used in all thesedifferent forms. As regards the peculiar Carabidse, Trachypachys and Systolosoma, we do not know enough of their mode of life to explain tlie structure : but as regards Silphomorpha and their allies we know that they are dwellers under bark, and we can feel sure that if this has been the case throughout a very long period of the ancestral record it must evolve a structural condition very different from that of the cursorial Carabidse. Omophron has a peculiar mode of life, concealing itself for long periods in fine wet sand, and we find that in it tlie articulation between the pro- and meso-thoraces is well fitted to guard against the entry of fine sand, although it would not be suflficient to keep water from gaining access to the interior.
The prothoracic stigma is placed high up, near the junction of the epimeron with the pronotum, and is protected by a hollow in the front edge of the uiesothoracic episternum ; the orifice is short and small, but the stigma is perfectly fitted for respiration, although it is not used for inspiration during the aquatic life of its possessor.
The Mesothoeax in the Dytiscidse might, when the under surface of the body is looked at, be supposed to be absent, for it is not exposed in front of the metasternum, but is, as it were, turned inwards, so as to be placed more or less at right angles with the longitudinal axis of the body ; the result of this is, that a kind of cavity or fissure, roofed above by the sides of the prothorax, exists in front of the metasternum, and in this hollow the flexed front and middle legs are packed during the process of swimming. This position of the mesosternum is however by no means invariable in the family, for though carried to an extreme extent in Hyphydrus, Hydrovatus, Cybister, and other higher forms, yet in the Vatellini, Sternopriscus and others the position differs but little from that of the Carabidte. The mesonotum has been hither- to but little studied although the visibihty or invisibility of the scutellum has largely determined the classification of the family. It offers however considerable modifica- tions which should not be neglected. The scutum appears in the form of two lateral lobes joined together along the middle line ; their base is nearly straight except that in the middle it is frequently (Dytiscus, Cybister, Graphoderes, &c.) intruded on by an angular projection from the middle of the front of the scutellum ; taken together the two lobes appear in front always more or less emarginate, being longest at the outer side, shortest at their point of junction in the mesial line ; this emargination may be very slight (Dytiscus bipustulatus No. 751 gen. Agabus e.g.,) or very deep (Dytiscus roeselii No. 1169, Cybister e.g.) In Noterus the base of the scutum is deeply transversely impressed, and the impression divided in the middle by a raised line, but in general the lobes are slightly convex, and without any definite impression. The mesonotum is very small in Eyphydrus, Laccophilus and
On Agnatic Carnivorous Coleoptera or Dytiscidce. 223
Hydrovatus, and the scatum is then very inconspicuous ; in Laccophilus especially it appears as a short band in front of the scutellum without any division into two lobes. I have already remarked that the hind margin of the scutum is impinged on in the middle by what appears to be an angular prolongation from the front of the scutellum ; this prolongation is however probably a distinct piece of the scutum, as it certainly is in the Carabidre, in some members of which family (Harpalus cahginosus e.g.) it appears as a large piece, nearly as large as the scutellum, reaching quite to the front of the mesonotum, and thus widely separating the two lobes of the scutum. The scutellum always comes to the surface of the body at the base of the suture of the elytra, but is very frequently invisible (Laccophilus, Noterides, Hydroporides and Methles) owing to its being covered by the middle of the base of the prothorax; but in the Macro-Dytiscidre the scutellum is always visible. In the groups where it is concealed it varies in form and size, thus in Noterus it forms a very short broad triangle the apex of which is directed to the hind part of the body ; in Laccophilus also it forms a short, broad triangle with rounded sides ; in Hyphydrus it forms an irregular four-sided figure, broadest behind, the hind margin straight and with projecting lateral angles ; in Hydrovatus it reaches its extreme diminution, and can only be detected as a minute transverse raised band : in the other Dytiscidse it has the form of a more or less elongate triangle, with the base in front attached to the scutum, but the sides forming a free edge, under which lock the anterior parts of the sutural edges of the wing cases. The post-scutellum is a transverse band on a lower level than the scutellum ; on it rests the base of the wing case when closed, and to its extremity is attached a small membrane ; this membrane is sometimes ciliate at its edges (Dytiscus, Cybister) and is always present in the Dytiscidae, and also in some Carabidse, though it has been supposed by Lacordaire (Introduction a r Entomologie, I., p. 347) to be peculiar to Dytiscus.
The most important piece of the mesosternum — the mesosternum proper or medisternum — consists of a middle portion, giving off above a lateral branch on each side, so that it may be roughly described as being somewhat T shaped ; the middle portion is the lowest part of the mesosternum, and has more or less the form (when disarticulated) ' of a short six-sided column, as seen in Dytiscus, Cybister, and most of the larger Dytiscidse, but in the case of such of the smaller Dytiscidaj as have only a very slight extension of the mesothorax in the longitudinal direction, the columnar appearance is quite lost (Laccophilus, Hydrovatus, Hyphydrus, &c.), and it is also but little conspicuous Avhen the depressed form of the insect causes the column to be very short (as in Copelatus). Like the prosternum proper, the medisternum appears to consist of two pieces placed one in front of the other, and distinguished by a very prominent line of junction. The front of the column is placed immediately behind the anterior coxse, and is on each side usually more or less hollowed for their accommodation, so that in fact the column bears two facettes in front which are sometimes separated by a channel (Cybister) ; in Dytiscus
224 On Aquatic Carnivorous Coleojjtcra or DijtiscidcB.
however the facettes are nearly or quite absent so that the front of the column is comparatively plane ; at its lowest point the column is very deeply channelled or impressed, to receive a protuberance of the posterior part of the prosternum, the sides of the channel being frequently divergent so that a kind of foi'k is formed (mesosternal fork) ; the front extremity of the fork is usually a slight angular prominence, and the hinder termination of the fork or channel, is sometimes also angularly pi'ominent (Dytiscus, Cybister, &c.); the floor (or roof) of the channel behind usually connects with the intercoxal process of the metasternum, and sometimes is prolonged backwards to effect this junction (D^^tiscus and many other Macro-Dytiscidas), while in other cases the metasternal process is sufficiently perfected to articulate completely wath the mesosternal fork without this latter being furnished with any backward prolongation (Cybister). Although the mesosternal fork is in the great majority of the Dytiscidse connected with the metasternum ; yet in Pelobius, in the Vatellini, iu a large portion of Hydroporini, and in Sternopriscus this is not the case, and the middle piece of the medisternuiii does not connect at all with the metasternum : this is of considerable importance, for in the Carabidae these pieces seem, so far as I know, to be always connected and are sometimes even soldered together (Anthia) ; the development of the pros- ternal process in the Dytlscidpe, is greater than in the Carabidre, while on the other hand the corresponding part of the mesosternum is smaller than in the Carabidae, though very highly modified in other resj^ects. It is well worthy of remark that even where the mesosternum fails to connect with the metasternumthere is frequently (Deronectes) a growth backwards from it tow'ards the nearest part of the meta- sternal process, as if a connexion were in process of evolution (for remarks on this, vide last part). Sometimes the apex of the medisternal column does not reach so far down as the front of the metasternum, and the connexion between it and the metasternum, if such exists at all, is comparatively imperfect (Eretes, Hyphydrus). It may also be remarked that the perfection of this connexion is to a considerable extent correlative with the distance between the intermediate coxse, as when these are very close together, and more especially when they are also prominent fi'om their cavities (Vatellini, Sternopriscus,) then the connexion is wanting, while when the coxse are well separated (Hydrovatus, Noterides, most Hydaticides) then the conne>:ion is perfect. The side of the middle column is distinct in most of the Macro-Dytiscidse (see Dytiscus, Meladema, Ilybius), but when the mesothorax has lost almost absolutely its extension in the longitudinal direction and become merely a diaphragm between the meta- and prothoraces (Laccophilus, Hydrovatus, Hyphydrus), then the column is so compressed that it becomes a mere lamina, bearing two facettes for the accommodation of the front coxae ; the two sides which form the back of the column are quite concealed so that they can only be viewed by disarticulating the mesothorax, they form a part of the anterior portion of the articular cavities for the middle coxae.
On Aquatic Carnivorous Coleoptera or Dytiscidoe. 225
The lateral wiiio-s of the medisternum are each very short in the lonoitudinal direction, and extend on each side so as to assist in forming the anterior portion of the middle coxal cavity ; they become each a little longer towards the outside of the coxa, where they terminate by the junction of the outer extremity with the lower portion of the epimeron ; throughout the family they are very constant showing no important variations in size and form.
The larger portion of the area of the mesosternum is formed by the episterna ; each of which is a large piece rising from the upper edge of the side-wing of the medisternum directly over the middle coxa ; its upper extremity is shorter (in the true longitudinal axis) than the lower, and rises higher than the other parts of the mesosternum and plays an important jJart in the articulation of the wing-case ; the inner margin of the episternum is the border of the large foramen affording the means of communication between the prothorax and afterbody; this inner margin has the appearance of being a distinct piece, articulating at its lower part with the front piece of the middle column of the medisternum (referred to above as being probably a distinct piece) ; at its upper extremity this inner margin expands and becomes hollowed, in a more or less evident manner, and thus affords space for the expansion or play of the prothoracic stigma which lies immediately in front of it. The episternum although forming so large a portion of the mesosternum is of little taxonomic importance, as it varies very little throughout the family, it is remarkably large in the Noterides, and in Hydro vatus a large part of its area is impressed for the accommodation of the middle femur. It is on the episternum that the torsion by which the mesosternum becomes diaphragm-like, is effected ; the front edge always preserves the natural direction, but the episternum is more or less gradually bent, so that the hind part of the mesosternum is placed something like at right angles to the front edge, which thus appears to form a very short neck.
The epimeron of the mesosternum is always smaller than the episternum, and its lower extremity always penetrates to the middle coxal cavity ; its inner (more correctly anterior) margin is closely connected with the episternum, the suture between the two being fine, but usually distinct, but in Bidessus and Sternopriscus it is most frequently very indistinct : the outer margin is applied to the front edge of the metathoracic episternum. The epimeron is always very short (in the truly longitudinal direction) at its lower part, but usually becomes broader as it reaches to the shoulder, and after its angle of junction with the humeral angle of the metathoracic episternum it again becomes narrower, so that it forms a rather irregular, narrow and elongate triangle ; in some cases, however, the epimeron becomes scarcely any broader at the metathoracic angle (Bidessus and Notei'ides) so that it may be described as linear, although in point of fact it always shows a well marked, sharply defined angle at its point of junction with the angle of the metathoracic episternum. In the Noterides the linear form of the epimeron is accompanied by a very large development of the episternum, but this is not the case
226 On Aquatic Carnivorous Coleoptera or Dytiscidve.
in Bidessus, where the episternum and epiraeron are each small in area ; the Noterides are therefore distinguished by the sublinear epimeron and the fact that this is very small in comparison with the episternum. Although the epimeron is always smaller than the e2:)isternum, yet in the higher forms it increases so much in size that it becomes almost equal to the episternum in ai-ea {vide Laccophilus, Acilius, and others). Tlie upper extremity of the epimeron, extends the whole length of the mesosternal flank, terminating, at the upper articulation of the wing- case, behind the extremity of the episternum ; its upper edge is grooved to receive the margin of the inflexed base of the wing-case.
There are some points specially characteristic of the mesothorax of the Dytiscidse, and a comparison of these with the Carabidte is of interest; they are — 1. The direction of the mesosternum ; 2. The magnitude of the anterior sternal pieces ; 3. The magnitude of the posterior piece ; 4. The separation of the middle coxee ; and 5. The penetration of the epimeron to the coxal cavity. 1. As regards the first of these, there does not exist to my knowledge any Carabid in which the direction of the mesosternum is so remarkably divergent from that of the meta- sternum as it is in most of the Dytiscidaj, and in the great majority of the members of the two families this difference is very considerable ; there are, however, some Dytiscidse — the Vatellini and Sternoprisciis, (and even a sjiecies of the genus Hydroporus, Dytiscus dorsalis No. 630) — in which the mesosternum has only to a comparatively slight extent assumed the diaphragmatic position it possesses in their allies ; and on the other hand in Systolosoma, Trachypachys and Cyclosomus of the Carabidfe, the direction of the mesosternum becomes to a great extent that of the DytiscidjB ; no absolute distinction exists therefore between the two families in this respect. 2. The medisternum, and the episternum are very much larger in the Carabidae than they are in the Dytiscidne (tlie head and prothorax must be separated from the afterbody before this can be appreciated). This distinction is remarkable and is no doubt correlative with, if not dependent on the fact that the presternum has a connexion with tliQ metasternum in the Dytiscidse, this being of course much facilitated by the abbreviation of the intervening mesosternum ; while as another explanation of the curtailment, we have the fact that the anterior part alluded to is prolonged to allow of a greater or less extent of rotation and nutation of the prothorax in the Carabidi:e, while in the Dytiscidse this power has been held in abeyance in consequence of the more supreme necessity for fixing and securing these parts to prevent the admission of water to the interior of the insect. Thus it is in the Carabidie where the head and prothorax are quite free and mobile that we find the greatest difference from the Dytiscidse in the point alluded to, while where there is a prosternal process in the Carabidge the medisternum and episternum become sliorter : Cyclosomus and Trachypachys, Systolosoma, with others of the fragmentary series of Carabidse (in which series it is that we find a prosternal process of not unfrequent occurrence) have therefore shorter anterior
On Aquatic Carnivorous Coleoptera or Dijtiscidoe. 227
pieces to the mesotliorax than have the other Carabidse. In none of them, however, do I find the abbreviation to be so considerable as it is in the Dytiscidte; in Systolosoma for instance (the only one of the three genera just mentioned that I have been able to dissect) the middle of the mesosternum is decidedly more elongate than it is even in the most Carabid like of the Dytiscidaa — Amphizoa. In describing the medisternum {ante page 223) I remarked that its central column appeared to consist of two distinct parts, and fui'ther that the inner or front margin of the episternum had also the appearance of being a distinct piece connecting with the front jjiece of the central column. In the Carabida3 these pieces are still more distinct, and it is in fact by the much larger size of these front pieces of the medisternum that the mesosternum of the Carabidaj differs from that of the Dytiscidse. As the structure of the mesosternum in the Carabidee has not so far as I am aware been thoroughly examined, I must for the present leave this subject, merely remarking that in Carabus catenulatus the middle of the mesosternum appears certainly composed of three parts super- posed one on the other. The third point in which the mesothorax of the Dytiscidas differs from that of the Carabidaj, is the greater development of the posterior side piece in the former group ; this is, however, by no means constant, for in the whole of the ti'ibe Noterides, as well as in some Hydroporides, of the Dytiscidie, the epimeron is as small and linear as it is in most of the Carabidse ; and in Silphomorpha of the Carabidse, the shape and size of the epimeron is similar to the Dytiscidse, (except that it does not penetrate to the coxal cavity). Fourth^ the rule in the Carabidas is that the intermediate coxse are very distinctly separated, but in the Dytiscidse they are more approximate ; this however is liable to numerous exceptions ; in the Pseudomorphides (and according to Lacordaire in the Ozoenides), the middle coxse are very approximate as in the Dytiscidse, and the Dytiscidse themselves differ considerably in this respect, thus although the coxse are con- tiguous in Vatellini, they are in Hydrocanthini and Suphisini as widely separated as they are in Carabidse ; and even in some of the higher forms — as the Hydaticini — the separation of the middle coxse is moderately broad. Fifth, in the vast mass of the Carabidie, the mesothoracic epimeron does not penetrate to the coxal cavity, whereas in the Dytiscidaj it invariably reaches the cavity ; but the first, or fragmentary, series of the Carabidse resemble the Dytiscidse in this respect, the epimeron reaching to the cavity as in the water-beetles.
Turning to the special points of approximation between the two families in respect of the mesosternum we find, first, that there is but little difference between the mesosternum of Systolosoma and Amphizoa, still the former is a little more Carabidiform than the latter in this respect ; and both differ but little from the Dytiscidse ; second, the mesosternum of Omophron is very similar to that of the Noterides ; third, although the Pseudomorphides resemble the Dytiscidse as regards the mesosternum, inasmuch as their coxse are approximate, and that sometimes
TBANS. ROT. DUB. SOC, N.S. VOL. II. 2 H
228 On Aquatic Carnivorous Coleoptera or VytiscidoB.
(Silphomorpha) its epimeron is Dytiscidiform in shape, yet the group differs widely from the Dytiscidce inasmuch as the anterior pieces are very large, and the epimeron does not penetrate to the coxal cavity. In Pelobius the mesothorax might pass as belonging either to Dytiscidse or Carabidae, the epinxeron being rather small and its superior metathoracic angle little developed, so that in this respect it is Carabidiform ; while as regards its direction it is indeterminate, and with respect to the size of the anterior pieces it is Dytiscid rather than Carabid. In the Haliplides too, the mesothorax fails to support their classification either with the Dytiscidse or Carabidse ; the epimeron is large, but quite different in form from any Dytiscidte, for it has no trace of any superior metathoracic angle, and its greatest length (in the true longitudinal direction) is not at the line of this angle, but near to the coxal cavity ; the anterior pieces are quite small, and different therefore from the Carabidse.
The Metasternum. The metanotum is of much larger area than the mesonotum, but in those forms in which the mesothoracic scutellum is least developed (Hydro- vatini, Hyphydrini, Laccophilini, and, but not so markedly, Hydroporini) it is much smaller than in those having the scutellum largely developed and of very firm con- sistency, as it is in the Macro-Dytiscidse generally. It is of complex structure, and it appears to me very difficult to determine accurately of how many pieces it is essentially composed, and to decide as to their homologies. It is of course covered entirely by the elytra, and although it is horny, it is thinner than are the exposed parts of the skeleton. It is traversed in the middle, by a deep and broad longi- tudinal depression, into which is received an incrassation on the basal portion of the inner face of the suture of the elytra. In the middle in front it is armed with a small angular vertical projection, which acts to some extent as a support to the free extremity of the mesothoracic scutellum, and which probably represents a portion of the prrescutum, which is otherwise absent : this projection is not present when the scutellum is rudimentary (Hyphydrini, Hydrovatini, Laccophilini), and is small in the Noterides, and Hydroporini ; in fact its development appears to be strictly correlative with the growth backwards of the scutellum. The larger part of the metanotum is occupied by the scutum, which is divided in a more or less irregular manner, longitudinally and transversely so as to form four pieces. The anterior pieces — one on each side — are shining and glabrous, and are especially reduced in size in the Hyphydrini, Hydrovatini and Laccojihilini, while in the Noterides these plates are small, and placed chiefly in the vertical instead of in the horizontal direction. Behind these two anterior plates there are placed the two posterior plates of the scutum ; these are not shining like the anterior ones, but are dull and frequently bear a few hairs. Behind the scutum, the scutellum is visible as a transverse linear band on each side, but in the middle it has a considerable extension in the anterior direction, by encroaching on the hind border of the posterior plates of the scutum. The post scutellum is placed behind the scutellum, and on a rather
On Aquatic Carnivorous Coleoptera or DytiscidcB. 229
lower level than it, it extends completely across the body as a horny band con- necting by its extremity with the metathoracic epimeron ; to its hind margin is attached the basal abdominal segment. In Hydrocanthus australasise this band is very peculiar, for instead of passing across the body in a straight line, it diverges backwards in the middle, so that between the scutellum and post scutellum there exists a large membranous space.
The metasternum consists of two side pieces on each side, and a middle piece — the metasternum proper ; this last is of very peculiar form, its hind margin is very firmly connected with the front of the hind coxa3, so that these are frequently supposed to be a portion of the metasternum ; in the middle in front this latter stretches forward and extends between the middle coxse, so forming an intercoxal process ; on each side of this are the middle coxae ; an inflexed portion of the metasternum forms in fact the posterior half of the coxal cavity ; outside the coxae the metasternum comes in contact with its episternum, and forms a curved or oblique suture in adaptation with this part ; its hind border is encroached on by the exten- sion forward of the hind coxse, and thus between the episternum and the coxa, the metasternum is more or less — frequently excessively abbreviated — so as to form a wing, called the lacinia or wing of the metasternum ; in the posterior direction the metasternum projects farther back at the expense of the coxae, so forming in the middle behind a more or less acute angle ; thus the metasternum is elongate along its middle, but extremely short externally on each side.
The intercoxal j^rocess in the majority of the Dytiscidis is connected in front with the fork of the metasternum, and when so connected it is marked in the prominent middle part with a groove or depression — the metasternal groove — for the reception of the apex of the prosternal process, which is usually lodged in it : the intercoxal process differs much in shape and in its various details, and is one of the best means of recognizing genera and species : when the mesosternal fork is not sufficiently elongate to reach the same plane as the front of the metasternum, then the apex of the metasternal process is curved or bent up, so as to enter the mesosternal fork (Dytiscus) or at least to touch it (Hyphydrus, Eretes) ; the nature of the intercoxal process seems chiefly to have been determined by the distance separating the middle coxae ; when the middle coxae are quite contiguous, and project from their cavities, then there is properly speaking no intercoxal process (Vatellini, Sterno- priscus) the middle of the metasternum showing merely a prominent angle behind, and in adaptation to the coxae ; where the middle cox^ are broadly separated (as in (Hydrovatini, Pachydrus) then the intercoxal process is correspondingly broad and short, and its connexion with the mesosternal fork is very intimate or exact, thus in Pachydrus each of its lateral angles projects and articulates with the raised angle of the extremity of the mesosternal fork, which in these cases is reduced, in the longitudinal direction, to a mere lamina, while the raised margin round the coxal cavity is continued inwards along the front of the intercoxal process ; in the
2 H 2
230 On Aquatic Carnivorous Coleoptera or Di/tiscidfe.
Hydrocanthini and the Suphisini, the intercoxal process is very large, being both elongate and broad, and is nearly truncate in front, in adapatation to the broad truncate extremity of the prosternal process, the two edges of these pieces fitting together at their line of junction in a most perfect manner : usually the intercoxal process is impressed to receive the extremity of the prosternal process, and the metasternal groove thus formed, coincides in shape with that of the point of the prostei'nal process ; when the intercoxal process is very broad, (Hydrocanthini, Suphisini, Hydrovatini, Pachydrus) then it bears no groove, for the prosternal process is brought very closely into adaptation with the intercoxal process ; and there is in these cases excessively little mobility of the prothorax, these being the forms in which this characteristic of the Dytiscidae has become almost absolute ; when the intercoxal process is very naiTow, (as in many Agabi, especially group 12; Eretes, &c.,) then the groove is elongate, narrow, and indefinite, consisting of a mere depression along the middle of the process ; and when the prosternal process is short, feeble, and acuminate, which usually occurs in such forms as have the middle coxse rather approximate, and the intercoxal process more or less curved upwards at its extremity (as in Hyphydrus, Coelambus, and other Hydi-oporini in which the mesosternal fork is not connected with the intercoxal process) then it can scarcely be said that there is any metasternal groove present ; while its maximum of per- fection is attained in the Cybistrini, where it is deep, with very definite sides, and perfectly adapted to the prosternal process, but yet when the point of the process is slightly lifted allows some movement of extension between the prothorax and after- body.
The central piece of the metasternum, is more or less produced behind between the front parts of the hind coxee, its sides usually meeting to form a more or less acute angle; in Hyphydrus however the posterior portion of the metasternum does not form an angle, but has a comparatively broad rounded extremity. From the front of the intercoxal process to its posterior angle the metasternum is usually elongate, but varies much in this respect and in Hydro vatus it is short and conse- quently its posterior angle is obtuse ; the greatest reduction of the middle of the metasternum occurs in the short globose Dytiscidse, and is exhibited by Suphis and Colpius, where its posterior part has scarcely any backward prolongation in the middle. Nearly always there is a fine longitudinal channel running along the middle of the metasternum, but this is frequently very obsolete. The most peculiar character exhibited by the middle of the metasternum, is that in the Noterini and Hydrocanthini, it is more or less raised or prominent, so as to form in conjunction with the prosternal process, and the middle portions of the hind coxae, a kind of broad flat keel, increasing in width as it progresses backwards; this peculiar structure reaches its greatest development in Hydrocanthus, and is without any jaarallel in other beetles. The laciniee or side wings of the metasternum are as it were compressed between the posterior coxse, lying behind them, and the episternum lying
On Aquatic Carnivorous Coleoptera or DytiscidcB. 231
in front of them, and their form therefore depends largely on the shape of these adjacent parts : when the hind coxa, has become very large it forms an arch in front and thus encroaches on the lacinia, which suffers a corresponding diminution in area, but as its external extremity is fixed to the outer side of the coxa, and this does not grow forward so much as the middle does, consequently the outer portion of the lacinia extends backwards outside the arch of the coxa as a more or less slender band according as the coxa is more or less extremely developed ; the extreme forms thus assumed are very remarkable, (see Hyphydrus, Eretes, Laccophilus) ; in the genus Agabus great variation exists as to the size of these lacinise, and in most other genera considerable diversity exists on this point. Usually the outer extremity of the lacinia terminates as a more or less acute point, but in the Thermonectini and Eretes the lacinia becomes a little broader near its apex, thus giving rise to a peculiar shape, which in Eretes where it is most extreme, may almost be described as clavate, the extremity of the- wing appearing obtuse or almost rounded. In the Noterides the wing of the metasternum assumes a peculiar form characteristic of the tribe, but as this as well as some other variations, is cor- relative with the shape and development of the hind coxa, it will be better spoken of when those organs are described. In Pelobius and Amphizoa, the metasternum has near its hind margin in the middle a transverse line or suture, which is not present in any other of the Dytiscidae, although it is a nearly constant character of the Carabidfe.
The metathoracic episternum is a large and conspicuous piece forming the shoulder of the metasternum ; it is nearly triangular in form, its front margin being in apposition with the hind margin of the mesothoracic epimeron ; the base of the triangle is adapted to the wing of the metasternum, and is curved in the Thermonectini and Eretes, Hyphydrus, Laccophilini, and in such other forms as have the hind coxae very largely developed, while in other groups it deviates but little from a straight line, till it is near the intermediate coxae, where it is always curved, sometimes (Hydaticus) abruptly so ; its inner angle penetrates to the middle coxal cavity in the Dytisci Complicati — a point of structure which has not yet been detected in any other Coleoptera — but in the Dytisci Fragmentati, it fails to reach the cavity, stopping short in the Vatellini at a considerable distance from the outside of the cavity, but in Laccophilus very nearly reaching to it. The posterior angle of the triangle extends as far backwards as, or even slightly farther than, the apex of the metasternal wing, and forms there a blunt, or rounded truncate angle, touching the margin of the elytron, or at a little distance from this : its third or humeral angle is covered by the wing-case ; the upper edge of the episternum just behind this angle shews a groove into which the inflexed edge of the epipleura fits. The anterior edge of the episternum is thickened so as to form a margin along its front, and when this margin is very large and much exposed as it is in Hydrovatus, and numerous others, it has
232 On Aquatic Carnivorous Coleo'ptera or DytiscidcB.
extremely the appearance of being a distinct piece of tlie sternum, whicli however is not the case. The upper or outer portion of the episternum is, at any rate in its anterior portion, covered by the wing-case, and when the part covered by the wino--case is large, as in Hydrovatus, then the posterior extremity appears very slender and acuminate, but on the elytra being taken off it will be seen that in such cases the termination of the episternum is really obtuse, the greater portion of its area being turned upwards at an angle with the exposed portion ; this covered portion of the episternum may readily be mistaken for the epimeron, more especially as it is of a different texture from the exposed part, and is traversed by a sort ot line, having somewhat the appearance of a suture, between the covered and exposed portions. The epimeron of the metathorax, unlike the episternum — is very indistinct, and might on a careless inspection, be supposed to be absent ; it is however always present and is placed on the upper posterior portion of the episternum, but extends farther backwards, than this piece does ; the suture between the episternum and epimeron is covered by the wing-case, and is often very fine and indistinct, but is alwa3's conspicuous when the part is properly cleaned and exposed ; the episternum towards its extremity becomes, as above stated, narrower, and thus trenches away from the edge of the wing-case, and an angle is thus formed between it and the upper and outer portion of the hind coxae, and into this angle the epimeron protrudes, so that at this point it approaches more to the under surface than it does elsewhere, and in many forms (especially in Eretes, Thermonectini, and Cybistrini,) the epimeron becomes visible, even when the wing-cases are closed, as an acutely angular projection at the upper and hinder angle of the apex of the episternum. Behind this point the epimeron again completely covered by the wing-case, proceeds farther backwards along the external or upper part of the hind coxa, till it terminates behind by connecting with the side of the basal segment of the abdomen or hind body; sometimes the connexion between these two parts is effected not by actual contact, but by means of an intervening membranous space (Acilius, Hydaticus, and many others, but in other cases (Hyphydrus, Eretes) there is complete contact between the horny portions of the two parts ; between the extreme degrees of separation (as seen for instance in Coptotomus) and the absolute contact of Hyphydrus and Eretes, there exist every grade of connexion ; thus in Agabus it would be almost equally correct to describe the two parts as separated by a membrane or as in contact. The upper portion of the epimeron becomes membranous, and connects with the pieces of the metanotum. In those cases where the middle coxae are widely separated, and there is no mobility of the pieces of the thorax (Hydrovatus, Hydrocanthus, Pachydrus) there is a most extreme and perfect adaptation ot the inflexed edge of the epipleura, to the side of the body, and in these cases the epimeron remains nearly entirely membranous, a mere strip of semicorneous matter, along its lower edge and its extremity, serving for the suture with the episternum
On Aquatic Carnivorous Coleoptera or Dytiscidce. 233
and hind coxa, while in Eretes where there exists considerable capacity for thoracic movement, and very little adaptation of the elytra to the sides of the body, the whole of the large epimeron is horny.
The more important of the characteristics of the metasternum in the Dytiscidae are, 1 , the peculiarly formed side wings ; 2, the intercoxal process grooved for the reception of the prosternal process, and 3, the absence of a transverse suture on its posterior part. None of these characters however are constant enough to serve as marks of distinction between the Dytiscidte and Carabidas ; the first of them is excessively variable in the Dytiscida ; it is chiefly dependent on the great develop- ment of the hind coxae, and in Amphizoa and Pelobius, and even in Colpius and Suphis, the form of the metasternum is quite Carabideous. The groove of the intercoxal process too is far from being constantly present in the Dytiscidse, it is absent as we have seen in all cases where the middle coxse are widely separated, and all that can be said is that it never occurs in the Carabidte, but is present in the great majority of the Dytiscidas. As regards the transverse suture in the metasternum, it occurs in the Dytiscidse only in Pelobius and Amphizoa, while on the other hand it becomes indistinct in some Carabidae, although I have not succeeded in finding any in which it is effaced.
The metasternum of the Haliplides has none of the peculiarities of the Dytiscidse ; the intercoxal process is not grooved ; the outline is completely Carabideous, and the transverse suture is present, though without dissection it is not easily apjireciated owing to a series of very large punctures placed on it.
The contiguity or amount of separation of the middle coxse in the carnivorous Coleoptera appears to me to be dependent rather on the metasternum than on the mesosternum. In the Carabidte these coxas are nearly ahvays very considerably separated by the intercoxal process, so that the one group of that family — the subfamily Ozsenidse — in which the coxse are unusually approximate is, according to Leconte, well distinguished by this character from all the rest of the family. In the Dytiscidse, greater variety in respect to the contiguity of the middle coxse is found, but still the character appears to me to be one of great importance ; indeed if I might exercise my imagination I would suppose that in the early history of the Dytiscidse the fact whether the middle coxse were well separated, or were approxi- mate, exercised a most important, or indeed predominant influence on the future mode of evolution. When they were distant this fact appears to me to have facilitated rapid co-adaptation of the contiguous parts for the purpose of protection and keeping out water from entering the body by the great fissure between the prothorax and the after-body by a process of mere placing together of accurately fitting surfaces ; and this involved complete disuse of any mobility or extension between these parts, and rendered sedentary habits of advantage, for in all movements there was the danger of these parts becoming separated, and it was necessary to keep the pieces of this great joint without any motion, and yet they
234 On Aquatic Carnivorotts ColeojJtera or Dytiscidce.
were not locked together by auy mechanical contrivance. But ia those cases where the coxse were more approximate, the accurate adaptation of the parts at this important spot was much less easily effected owing to the more complex faces to be co-adapted ; and the locking together of the pieces was effected by the more tardy but superior process of growth of the prosternal process to rest in a groove of the metasteruum, and by the fitting together of numerous less superficial parts and edges ; this plan probably placed the original developers of it for a period at a disadvantage, but permitted ultimately a development of the parts consistent with greater general activity and with a consequent general higher organization. In the cases where the coxse are absolutely contiguous (as in Vatellini, Andex and a few others) space was wanting to permit of the penetration of the prosternal process between the middle coxa3 and so the prothorax was fixed by a short process being received into the mesosternal fork, leaving a very incomplete joint; or (as in Sternopriscus the prosternal process grew backwards under the projecting middle coxjB (instead of between the coxa;) and so only connected with the metasternum in a very imperfect manner. However this may have been, 1 believe that the distance between the middle coxse has in each species been very constant during very long periods of its evolution, and may prove of much assistance in deciding as to the relationship of particular fossils to particular existing forms.
Hind Cox^. — Although tinily a part of the leg, yet the hind coxa in the Dytiscidse is developed to such an extent that it frequently forms a considerable portion of the external skeleton, and is therefore actually a part of the body rather than of the limb. It is from a taxouomical point of view, the most important part of the Dytiscidte, and its various modifications are very interesting on account of their great variety in detail and uniformit}- in more important particulars.
The pair of hind coxse extend all across the under surface of the body, and thus largely separate (when an undissected individual is looked at) the hind border of the metasternum from the ventral plates of the hind body. The coxa is completely destitute of mobility, for it is soldered in front, along its entire breadth to the metasternum, while behind it is accurately fitted to the basal ventral segment, and in the Bidessini is completely soldered thereto, so that in this group we find that the whole of the middle of the lower surface of the body — from the middle legs to the end of the third ventral segment, consists of a single, solid, hard piece ; in all other Dytiscidse, however, the ventral segments are not soldered to the coxsb ; althouoh the apposition between the two parts is so perfect that they look as if they were but one, yet really great mobility exists at this articulation. Each coxa articulates with five distinct pieces of the skeleton, viz., 1, its fellow coxa, 2, the metasternum proper, 3, the metasternal episternum, 4, the metasternal epimeron, and 5 the basal ventral segment. It consists of two distinct parts, viz., 1, the lamina interior, and 2, the lamina exterior. The internal lamina it is that carries the articulation of the swimming leg, and the two internal laminae are accurately
On Aquatic Carnivorous Coleojytera or Dytiscidie. 235
fitted together, and more or less strongly soldered together on the middle line of tlie body by a straight longitudinal suture ; a suture or impressed line runs near the outer border of the internal lamina, and proceeding forward, either reaches the anterior border of the coxa (as in most Agabi), or becomes obsolete in front ; in this latter case no real line of demarcation exists between the external and internal laminre in front of the part where this line stops short ; this line I call the coxal line, or speaking of the two together, coxal lines. The hinder part of the internal lamina is prominent, and presents a free edge over which the trochanter of the swimming leg plays, and which by its shape determines to some extent the kind of motion made by the swimming leg ; this is the coxal process, or coxal lobe. This process or lobe usually has, at any rate towards its termination, a kind of external margin or border, marked off by the terminal portion of the coxal line ; this is the coxal border. The hinder edge of the coxal process is very frequently marked by a short closed suture (see Dytiscus) or a small notch or emargination ; this I have called the coxal notch. The external division or lamina of the hind coxa is always larger than the other, sometimes enormously larger ; its front border proceeding from the middle is directed outwards and at the same time more or less forwards, l)ut before reaching the side, turns more or less abruptly backwards, so as to form a sort of curve or arch of very diverse forms ; it is soldered to the metasternum, but it passes outwards beyond this latter behind the posterior point of the epister- num, thus reaching the margin of the epipleura ; it does not, however, there terminate but (as is seen on taking off the wing-case) is reflexed upwards passing behind the epimeron and terminating externally, in an either angulated or rounded form, by articulation with the outside of the post-scutellum, by the intervention of a small piece considered by Schioedte (Dan. El. p. 377) to be a trochantin. The hind margin of the exterior lamina (as seen where the specimen is entire) is likewise directed somewhat towards the front, and near the epipleura forms a considerable curve or sweep, which is continued under the wing-case, to the outside extremity of the coxa ; this apparent hind margin is not however the real hind margin, for on dissection it is seen that the coxa is to a large extent reflected in a very round curve upwards to the interior of the insect ; this reflected and concealed portion of the coxa may be called its abdominal portion, and in Hyphydrus is very distinctly marked off along the whole breadth of the coxa by a raised line or margin. The -exterior lamina of the coxa differs greatly in size and shape, and its variations offer one of the best modes of distinguishing the genera and species. The changes of form occur on its anterior portion. Tlrus in Pelobius and Amphizoa where the Dytiscid coxa is seen in its most rudimentary condition, the metasterno-coxal suture proceeds directly outwards, without stretching forwards at all, so that the two sutures together form a nearly straight line extending all across the under surface : we find the other extreme in the genus Eretes, where the metasterno-coxal suture proceeds almost directly forwards, thus the sutures of the two sides at their
TBAItS. BOY. DUB. SOC, N.S, VOL, n. SI
23t) On Aquatic Carnivorous Coleoptcra or DytiscidcB.
point of connexion form a very acute angle, and proceeding forwards like the branches of a letter V diverge but little till they have become quite near to the middle coxa when they are rather abruptly turned outwards. In Hydrovatus these sutures of the two sides diverge from one another only at a very obtuse angle, and therefore are directed but little forwards at first, but when rather near to the out- side of the body, the direction becomes more abruptly forwards, but only to be almost immediately even more abruptly bent back : a similar form of the coxa is pretty general in the Hydroporini. The anterior border of the coxa of the Dytiscidpe thus forms in front an arch which is very various in its form, in Eretes as we have seen it is a broad, very flat arch ; while in Hydrovatus a naiTow abrupt curve placed near the outside of the body is formed. Yaiious intermediates between these extremely different forms occur ; and it is only in the Noterides that we find a really distinct form of the coxa, owing to the fact that this part attains its greatest extension forwards near to the middle of the body, and then becomes shorter towards the side. In the Cjibistrini the coxa assumes in front a shape characteristic of the group, its greatest anterior extension being gained near the outside of the body, while external to this point it so abruptly retreats as to form almost a right angle. In size tlie external lamina varies even more than it does in foi'm; it is, compared with the average of other Coleoptera, always large, and within the bounds of the family is least in Pelobius and Amjihizoa; in certain species of Agabus (Dytiscus uliginosus. No. 694, Agabus maderensis No. 666), it IS not greatly larger than in Amphizoa, and with every gradation of growth in various species and genera, reaches a truly enormous size in Hyphydi'us, Eretes^ Acilius and Coptotomus.
Though the external lamina is of much interest and importance, yet the internal lamina on account of its being the seat of the articulation of the portion of the leg used for swimming is much more important, and much more complex and varied. The base of each trochanter is placed in a kind of box, and the larger the opening of the box the greater is the range of motion of the limb, and the more powerful its sweep, and in the higher forms of the Dytiscida? the articulation is constructed so as to allow the leg to make a sweep extending round the complete half circum- ference of a circle, and yet the articulation is so well constructed that no water can obtain entrance by it to the interior of the body where the muscles ai'e situated. The socket is formed above by an arched plate, which appears to be merely a prolongation of the abdominal portion of the external lamina of the coxa, and below by the coxal lobe or process, there being left between these two parts a large circularly transverse cleft looking backwards and laterally ; the internal Avail is concealed owing to its being closely connected with its fellow of the opposite side of the body ; in fact in all the higher forms of the family, there seems to be only a single opening for the two legs — in other words the two transverse clefts appear joined into one — but when a dissection is made, it is seen that they
On Aquatic Carnivorous Coleoptera or Dytiscidce. 237
are separated by a vertical wall, or plate ; this wall ascends from the common suture between the two coxte, and at its termination, spreads out above into a kind of winged or bifid process which completely fills up a gap that would otherwise exist between the arched roofs of the two cavities. The articular portions of the coxa project backw^ards, and on each side of the projecting portion there is a hollow or axilla which permits the base of the swimming leg to be rotated forwards till the front edge of the femur can quite attain the longitudinal middle line of the body. In the higher and larger forms of the Dytiscidre which I am now describing, the articular cavities are entirely concealed by the coxal processes, which project from the surface more or less slightly backwards, and are thus very consjiicuous on the under surface of the body, the swimming legs protruding as it were from the hind part of the projection, and being so close together that their inner margins touch one another at the point of articulation. This structure, with modifications in some details such as the size and form of the coxal lobes, prevails throughout the Macro-Dytiscidte, without any exception. But in the smaller Dytiscidse the articulation of the swimming legs is of a different and much less uniform nature. If a Hyphydrus be looked at — and with this object one of the two swimming legs should be carefully disarticulated — it will be seen that the swimming legs do not project from behind and above any prominent coxal processes, and that they are not contiguous, but are separated by a considerable space, and if the leg has been carefully disarticulated, without damage to the articular cavity it will be seen that this is a circular orifice, completely exposed and not concealed by any projecting lobe, and a farther examination renders it evident that the space separating the articulations of the two sides of the body, consists of the coxal processes. Here then we have the articular orifice exposed and placed outside of the coxal process, separated from its fellow and circular in form, instead of being placed above the coxal process, concealed by it, and contiguous with its fellow ; on separating the hind body from the coxa it is farther seen that the upper portion* of the articular box projects further back than the coxal processes, these latter, being adpressed to, and soldered to the former : thus the coxal processes, are not prominent from the level of the base of the abdomen, and there is no axilla formed to permit the flexion of the leg forwards : indeed the articular cavity being circular and exposed permits the leg to be rotated forwards even farther and more freely than in the Macro-Dytiscidae. This is the type of structure which admits of the greatest amount of motion of rotation for the swimming leg and it attains its maximum of develop- ment in the New-World genus Pachydrus, where it is accompanied by a consolidation of the coxa with the abdominal segments. Many of the Hydroporini have the articular cavities constructed in a manner that is intermediate between that just described, and that of the Macro-Dytiscidse ; as a good instance of an intermediate
* I have in my descriptions occasionally spoken of this upper portion of the articular cavity, when it is visible under the name " pyxal process."
2 1?
238 On Aquatic Carnivorous Coleoptera or Dytiscidce.
condition, Hydroporus gigas No. 427 (Chostonectes) may be mentioned. In Pelobius and the Hydrovatini we have another form of articulation of ihe swimming leg -. inasmuch as the coxal cavities are rather widelj^ separated, but yet are concealed by a coxal lobe; the inner termination of the articular cavity appears therefore in these cases in the form of an opening or excision on the hind margin of the internal lamina. I have already spoken of a mark which commonly exists on the hind margin of the internal lamina, either as a slight notch or interruption of the outline of the coxal process (C^^bister, Ilybius, &c.,) or as a closed shoi't suture (Dytiscus). This coxal notch is remarkably persistent, and indeed thougli it appears to be absent in many forms (Acilius e.g.) yet on careful examination I have alwaj's found some traces of it, it exists also in the Carabidse, and is perhaps an indication that this part of the coxa is formed by the coalescence of two distinct pieces ; its persistence is at any rate very remarkable, and is clearly an indication of some formerly different structural condition of these parts. I had thought at first that the articular opening existing in Pelobius and Hydrovatini was merely an enlargement of this coxal notch but such is not the case, for in Pelobius the coxal notch may be seen existing on the coxal process, just outside the articular opening. In the Noterides the coxal lobes are of large size and peculiar form, reaching furthest backwards at their outer angle, which is thus more or less acute, they are greatly prominent from the rest of the under surface, and thus form a deep elongate axilla, while the articular cavities them- selves are completely concealed ; in respect therefore of the development of the coxal processes, and concealment of the articular cavities, the Hyphydrini and the Koterides stand as the op^josite extremes of the family. In Amphizoa the articular cavities are widely sejDarated, and are concealed, but their inner termination is distinctly visible on the hind margin of the coxal process, the coxal notch being placed immediately outside of this opening : this differs from the articulation in Pelobius, inasmuch as the termination of the articular opening is not prolonged forwards, and thus forms a shallow emargination instead of a deep excision, if we suppose the space separating the two cavities in Amphizoa to be absent, the structure would not then differ materially from what exists in the Macro-Dytiscidaj.
The structure of the hind coxa in the Dytiscidse may be summed up as differing from that of the Carabidce in the following points ; 1. The large size or area of the external lamina ; 2. The arched anterior border of this part ; 3. The accurate adap- tation and soldering of the two internal laminje together by a straight and perfect suture; and 4. The contiguity of the articular cavities. As regards the first of these points, it is certain that so far as the large majority of the Dytiscida3 are concerned, the greatly increased size of the outer lamina of the hind coxa is very remarkable and is one of the modifications of structure most characteristic of the family : it is however only a question of degree, and varies extremely in the different groups, so that Amphizoa and Pelobius, in which forms it is smallest, appear rather to be Carabidaj than Dytiscidee in respect of this peculiarity : I am not aware however of
On Aquatic Carnivorous Coleoptera or DytiscidcB. 239
any Carabidaj, in which the area of the external lamina of the hind coxa is anything like as large as it is even in Pelobius and Amphizoa. In Trachypach3's and Systolo- soma of the Carabidne, the hind coxa is large, but it is made so by the increase in size of the internal rather than of the external lamina. In Silphomorpha the in- ternal lamina is very small, and the external one correspondingly large, thus making a slight approximation to the Dytiscidse. [Second, As regards the arched anterior border of the hind coxa in the Dytiscidse ; such a character is I believe absolutely unknown in the Carabid£e,but unfortunately it is not always presentin the Dytiscidse, being absent in Pelobius, Amphizoa, and Colpius, and very nearly absent in Suphis. Third, The accurate coadaptation of the two internal laminae at their junction on the middle line is quite characteristic of the Dytiscidse and is always present ; it is however approximated by some Carabidse, viz., the Pseudomorphides, Trachypachys and Sj'stolosoma ; but the only one of these that really makes any near approach in this respect to the Dytiscidte is Trachypachys. Fourth, As regards the contiguity of the articular cavities, this is very conspicuous and characteristic in the Macro- Dytiscidte, but in many other Dytiscidse it is not present (Pelobius, Hydrovatini, Hyphydrini, Bidessini, Colpius), indeed a slight separation of the cavities is so common in the lower forms of the various groups, as to strongly suggest the idea that all the species of Dytiscidaj have had ancestors with separated posterior coxal cavities like the Carabidae.
The Haliplides show not the least approach to tlie Dytiscidfe in the structure of the hind coxae ; on the contrary they possess a peculiar development, which is not approximated by any Carabidse or Dytiscidae, of these parts.
HiND-BODY, or Abdomex. — The dorsal plates of the hind body are eight in number, and they differ but little from one another in length : they are membranous in tex- ture, but usually dark in colour ; the seventh and eighth are a little thicker than the others, and have thus a leathery consistence, and they are also dull and more or less punctate, especially the eighth, while those in front of them are shining : in Dytiscus the two basal segments bear a large quantity of very fine, elongate hair, and other genera show sometimes a similar development though to a less extent. In Hydro- vatus the dorsal plates are very thin and delicate, and pallid in colour, and the apical one differs but little from the others. The basal segment is attached to the hind margin of the metanotum, and each segment is attached by its sides, b}' the intervention of a very delicate membrane, to the harder side pieces of the body. The metathoracic stigma is placed at the hind margin of the metathorax, at the side of the body, and may be either small and inconspicuous (Cybister), or elongate in the transverse direction (Dytiscus) : there are seven pairs of true abdominal stigmata ; the first abdominal plate is without a stigma, but each of the other plates bears at the outside a stigma placed in its delicate membranous border, the stigma of the eighth or terminal segment, is usually placed quite at its front edge, but in Cybister is placed near its hind margin. The stigmata vary greatly in their size and develop-
240 On Aquatic Carnivorous Coleoj^tera or Dytiscidce.
ruent and are at their maximum in Dytiscus, where all are large, and those placed on the two apical segments are, contrary to the usual rule, larger than the others, and are of a transversely elongate, elliptical form. In Cybister and Megadytes the abdominal stigmata are all small, and the quite small apical one, passes through the segment in the form of a tunnel, instead of opening directly through it as it does in others of the family ; moreover in the two genera just named, the terminal segment appears to be much more retractile and mobile than it is in other forms. The side pieces of the hind body consist of a hard membrane, becoming externally harder and corneous, and attached to the side of the turned up edge of the ventral plates ; the side piece of the basal segment is marked in the Colymbetini and Dytiscini with transverse rugse, which do not occur in any other Dytiscidse, and are probably of some assistance in the process of respiration.
The ventral plates or segments, are six in number, and are hard and corneous ; they are transversely arched and each one has its outer margin turned upwards, and even somewhat inwards, and this part is marked off from the rest of the surface by a raised margin, so that the upjser edges of the ventral plates look at first sight as if they did not belong to this part of the body, but were rather the abdominal side pieces ; the margin giving them this appearance is however nothing but a raised carina for the accommodation of the edges of the elytra, a similar margin being developed, on the outside of the hind coxa and even along the side pieces of the metasternum. The basal segment is very much modified in form to accurately adapt it to the hind coxa, and for this purpose its middle appears to be completely cut away, so that in an undivided insect, the first visible ventral segment appears to be separated into two pieces placed one on each side and separated in the middle by a considerable interval ; but on dissecting off the hind body it is seen that the two pieces are connected together by a slender isthmus, but that this middle piece or isthmus is concealed by being turned upwards at right-angles to the rest of the plate. The basal segment carries on its upper face (and therefore as it were in the interior of the body) a transverse corneous partition, extending all across the hind body, and adapted to the reflexed postei'ior portion of the hind coxa ; this concealed transverse partition is the true first ventral plate, the basal plate just spoken of being really a portion of the second segment : it results from this arrangement that the first dorsal plate is attached to what is truly the second ventral plate, the second dorsal plate to the third ventral, and so on ; but this nomenclature is not in use and the first visible ventral segment or true second plate is called the first plate. The second ventral plate is more or less slightly emarginate in the middle in front, by being adapted to the projecting internal laminte of the coxfe ; in some cases this emargination is but slight (Cybister, Dytiscus) but in others Dytiscus duodecim-pus- tulatus No. 462 e.g.) it extends nearly to the hind margin of the segment, so that in such a case if the ventral plates are counted along the middle line they appear to be only four in number. The sutures between the first and second, and
On Aquatic Carnivorous Coleoptera or DijtiscidcB. 241
between the second and third segments are more or less fine and indistinct, these segments being in fact soldered together and immovable, but the remaining ventral sutures are distinct, the plates being mobile : the fourth and fifth ventral plates are shorter than the others, but usually not very greatly so ; the sixth or last plate is more elongate, more or less obtusely pointed, and with its hind edge more or less finely margined — very finely in Hydroporides, coarsely in Cybister ; in Ilybius this plate shows a considerable difference of form according to the sex, but this is a rare exception. The outer or upper portion of the ventral segments, placed under the wing-cases, and marked off as I have already described by a raised carina, varies a good deal in its width, especially on the last three segments ; It is broad in Hyphoporus and in the convex forms, such as Hydrovatus, and Hydro- canthus, and it is very narrow in Dytiscus ; in the Colymbetini this upper piece of the fourth and fifth segments is much narrower than it is in the Agabini.
No important or constant character can be pointed out as distinctive of the Dytiscidte in opposition to tlie Carabidre as regards either the structure or con- nexions of the hind body. In the Dytiscidre owing to the great development of the hind coxte, the abdominal segments in the middle of the body are to all ajDpearance widely separated from the metasternum, whereas in the Carabidse it is quite frequently the case that the second ventral plate of the hind body, touches in the middle of the body the point of the metasternum ; but there are numerous Carabidse in which this is not the case (see especially Pseudomorphides and Trachypachys) .; the fact is not of any importance otherwise than as indicating the constant and complete apposition and union of the two internal lamince of the coxas in the Dytiscidse.
Recollecting the aquatic life of the Dytiscidte, and their very peculiar method of obtaining a supply of air, we should expect to find some notable character in the breathing orifices or stigmata. Such is not the case however, and it is only in the few species constituting the group Dytiscini, that we meet with any peculiar development of the stigmata ; while in the larger portion of the family the stigmata show no character by which they can be distinguished from those of the Carabidee.
Elytra or Wing-cases. — The elytra in the Dytiscidoe play a very important part, as it is by their means that water is excluded from all the dorsal portion of the body behind the pronotum. They are always hard and never abbi'eviate, and they fit so accurately to one another along their suture, and at the sides of the body, that the insect is enabled to carry about under them in the water a supply of air for respiratory purposes. They are never soldered together at the suture as is so frequently the case in the Carabidse, and they are kept in the closed position by an extremely beautiful combination of adaptations of various parts ; they lock together at the suture by a kind of joint similar to that called a rabit by joiners : a raised line is developed on their inner face which serves as a stop for the upturned edges of the ventral plates to repose against, while the inflexed epipleura rests its edge
242 On Aquatic Carnivorous Colcoptera or iJytiscidce.
ou a raised carina developed near the borders of the ventral plates and the outside of the hind coxa, and at the shoulder is received in a groove on the upper edge of tho metathoracic episternum; the base rests directly on the vei'tical mesosternum, and the sutural angle is shaped so as to fit under the edge of the raised scutellum ; by these means accurate adaptation of the wing-cases to the after body is secured, and then the hind margin of the pronotum rests on the base of the elytra, while the prothorax itself is firmly kept in its position by its prosternal process being locked in the metasternal groove ; t'lis latter part in f;\ct plays the part of the keystone of the arch.
Each elytron is articulated to the mesothorax by a projection from the base of the former about half way between the suture and outer margin ; external to the articular point the front margin is turned downwards, till it reaches the humeral angle, which is usually more or less prominent and nearly rectangular ; behind the shoulder the margin continues deflexed, and persists till towards the hind part, gradually becoming narrower, it ceases altogether at some distance before the hind sutural angle ; this deflexed portion of the wing-case is called the epipleura, its lower (or inner) edge is bordered by a very fine raised margin, and its junction with the upper surface is more or less distinctly marked by a raised edge, which forms the external outline of the body (after the prothorax). The epipleura attains its maximum of development so far as this family is concerned in some species of Megadytes (Cybister costalis, e.g.) while its minimum is found in Eretes and Ccelam- bus ; and in many of the family the epipleura becomes extremely slender behind the middle (Agabus, Ilybius, &c.) In some forms of Hydroporides (Hydrovatini, Hyphidrini, Ccelambus, &c.) there is more or less conspicuously marked off at the shoulder of the e^Dipleura a triangular space, which I have called the genicular or humeral area ; this receives the knees of the front and middle legs and permits these parts to be very closely packed away when not in use. The epipleura never extends so far as the hinder sutural angle, indeed the extremities of the wing-cases always terminate as a thin edge, and thus permit of the protrusion of the hind margin of the terminal segment of the hind body ; a supply of air is obtained during life by the insect separating these two parts which are protruded at the surface of the water, and are closed when under water by accurate apposition ; the inner face oi the elytra at this apical part is frequently pubescent (Ilybius, Cybister, Dytiscus) • the terminal portion of the wing-case is frequently slightly twisted or arched (Ilybius) and sometimes also slightly truncate (Lancetes, Dytiscus) : the sutural angle varies from being about rectangular, (Dytiscus) to very acute (Hj^drocanthus).
On the inner face of the wing-case near the outer margin there frequently exists a ridge parallel to the margin, this is frequently absent (Eretes) or but slightly developed (Rhantus, Acilius) : this elytral ridge is wanting on the front part of the wing-case, only commencing about the middle or behind it, and usually does not extend to the extremity, but disappears a considerable distance before that ; in
On Aquatic Carnivorous Coteoptcra or Dijliscidce. 243
A-gabus however it usually extends to the apex. In Hyphydrus the elytral ridge
shows a very peculiar structure ; a little in front of the extremity there is an
elongate prominent tongue, which projects towards the edge of the wing-case and
even a little beyond it, leaving a small space between it and the inner face of the
wing-case, and into tlie deep groove so formed fits the upper portion of the ventral
plate ; as the result of this the two wing-cases at the extremity are so firmly locked
together, that they can only be separated by first lifting them up from the body
till the extremity of the ligula is free : a similar but smaller ligula is seen in
Ccelambus, and in some other genera of Hydroporini it is found in a still less
perfected state. In Queda and Pachydrus the elytral ridge attains its greatest
development ; it is elongate, and for nearly its whole length forms with the face of
the elytron a kind of groove into which are received the edges of the vertical
segments ; near its extremity it becomes more prominent and thickened, and waved
in such a manner as almost to form a broad double ligula, the elytra are thus locked
together in an extremely firm manner. In Pelobius there exists two structui'es of
an unique character on the inner face of the wing-case ; one is a raised longitudinal
band at the extremity near the suture, on each elytron ; these bands are marked
with transverse strife, and are a part of an apparatus for producing sound. The
second is a large protuberance near the outer margin, just in front of the middle ;
this protuberance forms a cavity into which is received a horny process on the upper
edge of the basal ventral segment, and like the ligula in Hyphydrus serves the
purpose of very firmly locking together the two wing-cases.
WixGS. — Wings are always present in the Dytiscidse, and are well developed and large, except in a very few species (Colymbetes bifarius No. 75 7 e. g.) where they are rudimentary. They are folded in a complex manner, and placed in close adaptation to the inner surface of the wing-cases, not touching the body, and thus when the elytra are more than usually diaphanous (as in many Laccophili) the veins of the wings may be slightly visible through the wing-cases, giving the latter the appearance of being marked or coloured in an unusual manner. There are two different folds to adapt them to the size of the wing-cases ; one of these is a simple lapping over of the basal portion, but the other is a very complicated longitudinal and transverse fold and is situated on the front part of the wing just beyond the middle, and is exactly similar to the corresponding fold of the wing in the Carabidi« and Cicindelidoe. The costal and subcostal nervures are close together, and form the front border of the wing, the former of them is continued to the apex of the wing, and it is by traction on this costal nervure that the wing is unfolded, as may be readily ascertained by cutting off the wing of a freshly killed specimen, and then pulling the base of this nervure with a pair of forceps ; the subcostal nervure on the other hand is continued only as far as the transverse fold (or carpa) of the wing, when it is turned backwards, so as to leave an irregular cell or stigma immediately behind the costa ; this subcostal nervure is, when
TRANS. ROY. DUB. SOC, N.S., VOL. II. ^ ^
244 On Aquatic Carnivorous Coleoptera or Dytiscidce.
traction is exercised on its base, of assistance in folding the wing. The median nervure is a broad one, and runs from the base