594 



NATURE 



{Oct. 1 6, 1879 



the lateral gill-slits are still visible, are even larger as compared 

 with the parent animal than is the case in Ctecilia cotmpressicauda. 

 According to this, one might almost believe that such a relation 

 in size is universal in the embryo Caecilians. On the other 

 hand, one might conjecture that a development with external 

 vehicular gills, which obtains only exceptionally amongst the 

 Anura, as in Ofisthoddphys and No/olrcma, occurred more com- 

 monly amongst the Cfecilians. The few observations however 

 which have as yet been made on the other species of Coecilians 

 do not confirm this. Thus A. Dumeril, in a young C. oxynira, 

 50 millims. long, has found on each side of the neck a branchial 

 cleft, which it is true lies somewhat higher than in E. ^lutinosiim, 

 but still proves that in this species no external vesicular gills 

 are developed {Mem, Soc. Sc. Nat. Cherbourg, ix., taf., I Fig. 8). 

 Further, Prof. Mobius on the occasir)n of his late vi-it to the 

 Seychelles, brought back several examples of C. rostrata (Cuv.), 

 of very varying sizes (from 35 to 240 millims.) which have neither 

 branchial clefts, nor a swimming-tail, nor do they show the scars 

 on the neck that accompany the vesicular gills. All this leads 

 one to conclude that the development of the Cjecilians, like that 

 of the Anura, goes on in very various ways, and that in this field 

 too important discoveries are still to be made. It cannot therefore 

 be too deeply impressed on naturalists who visit tropical countries 

 where Caecilians are found, to give to this .subject their especial 

 attention." 



PHILOSOPHY OF THE PUPATION OF SOME 

 BUTTERFLIES^ 



'T'HE comparatively sudden transitions from one state to 

 "^ another in insects, have always excited the keenest 

 interest. The change from larva to chrysalis in those butter- 

 flies known as snspensi, and which in the chrysalis state hang 

 from the tip of the body, has, perhaps, been looked upon as the 

 most wonderful. The preliminary acts in the performance have 

 been pretty well observed and described by various authors 

 since the days of Vallisnieri. The larva hangs by the anal end, 

 turning up the anterior part of the body in a more or less 

 complete curve, and the skin finally splits from the head to 

 the front edge of the metathoracic joint, and is worked back 

 in a shrivelled mass towards the point of attachment. Now 

 comes the critical feat which has most puzzled naturalists, viz., 

 the independent attachment of the chrysalis and the withdrawal 

 from, and the getting rid of, the larval skin which such attachment 

 implies. 



Reaumur explained it in 1734 by the clutching of the larval 

 skin between alternate sutures of the soft joints of the chrysalis ; 

 and his happy and circumspect account from observations made 

 on Vanessa urlica has formed the basis for subsequent accounts ; 

 no one obtained a deeper insight into the philosophy of the act 

 until, some two years since, Dr. J. A. Osborne, of Milford, 

 England, discovered that a distinct membrane is concerned in it. 

 In casual observations of the process I had long become con- 

 vinced that the popular accounts were crude and inaccurate, and 

 I had preserved specimens in the act of transforming, for future 

 study ; but the philosophy of the change cannot be satisfactorily 

 made out from alcoholic specimens alone, nor from the study of 

 one species. The present paper is based on observations made 

 on species belonging to more than a dozen genera, the conclusions 

 having been partially presented last June to the Philosophical 

 Society of Washington. 



The body of the larva is composed (exclusive of the head) of 

 twelve segments or joints, and a sub-Joint. It is with this sub- 

 joint that we have here to deal, for to it beneath the rectum are 

 appended the anal pro-legs, and above this is the anal plate. 



If we carefully examine the anal plates of the larva; of the 

 true susfensi, we shall find that while they differ in form they 

 have one feature in common, viz., the being furnished dorsally 

 and posteriorly with numerous short spines and points generally 

 retrorse, or so placed that the larva can make use of them in 

 suspending. These special spines on the anal plate are only 

 fully developed after the last larval moult, being more or less 

 obsolete in the earlier stages, and they are also under muscular 

 control. Even in the succincti, where, as a rule, the anal plate 

 is not specialised, spines are, nevertheless, sparsely found, 

 especially on the border. 



All writers whom I have consulted speak of the larval suspension 

 being due to the entanglement of the hooks of the anal pro-legs 



' Abstract of a p.iper read before the American Association ior the 

 Advancement of Science, by Prof. C. V. Riley.^ 



in the silk, and do not mention the use of the anal plate, for 

 which the hillock of silk is sometimes spun in special form.' 



The normal form may be likened to that of an inverted settee, 

 or shoe, or to a ships-knee, and one of the most interesting acts 

 of the larva, preliminary to suspension, is the bending and 

 working of the anal parts in order to fasten the back of the plate 

 to the inside of the back of the settee, while the crotchets of the 

 le^s are entangled in the more flattened position or -eat. In 

 some cases (as in Danais) the hillock of silk is more elongate, 

 and the spines of the truncate plate mostly occur around ihe liwer 

 margin and even beneath it, so that in fa tening them the larva 

 seems to he drawing the silk up the rectum. In other cases (as 

 in EuptoUla) the plate, in addition to the spines, has a prominent 

 tubercle on each anterior outer border well calculated to lock 

 securely into the silk. After suspension, and as the fluids gravi. 

 tate anteriorly, the silken hillock becomes more conical (the 

 threads being loosely spun and elastic) and the hooks both of the 

 plate and the pro-legs hang more loosely from it. 



In the final getting rid of the larval skin and attachment of 

 the chrysalis there are concerned — 



I. Certain factors belonging to the larva and cast off with its 

 skin. 2. Those belonging to the chrysalis ; and to intelligibly 

 explain the process it is necessary to more fully characterise and 

 homologise these parts than has hitherto been done. 



In the former category, in addition to the natural adhesiveness 

 of the moist, mucous, and membranous corium,'' there are three 

 physiological factors concerned: (i) the tracheal ligament, t.-c 

 the shed trachere from the last or ninth pair of spiracles w hich 

 uniformly become blind or obsolete in the chrysalis ; (2) the 

 rectal ligament or shed intestinal canal ; (3) the Osborne or 

 retaining membrane {mevibrana retinens), which is but a stretched 

 part of the membranous coriam that accumulates aroui.d the 

 rectum and in the anal pro-legs. 



In the second category we have the structural features of the 

 chrysalis. These are, first, the cretnaster proper, which is the 

 homologue of the anal plate of the larva, and the form of « hich 

 is foreshadowed in that of said anal plate. This cremaster 

 assumes a great variety of different forms, but in general may 

 be said to be a tapering piece more or less incurved ventrally, 

 and having the ventral and dorsal margin- thickened or ridged, 

 and these ridges may be respectively called the ventral and the 

 dorsal cremastral ridges. This cremaster is surmounted at the 

 apex and sometimes along the ventral ridges by what maybe 

 called the cremastral hook-pad, thickly studded wilb m-nute but 

 stout hooks, which are sometimes comjjound or furnished w;tb 

 barbs very much as are some of our fishing-hook-, and which 

 are most admirably adapted to the purpose for which they .-.re 

 intended. 



Secondly, we have the sustainers {sustentores), two projeciicns 

 which homologise with the soles (planttz:) of the anal , ro-le<;s, 

 and which take on various forms, but are alwa)s directed for- 

 ward, so as to easily catch hold of the retaining uiemhraiie. In 

 the yellow butterflies (as Callydrias, Terias, Colias), where the 

 body of the chrysalis is so thrown back that mere projecting 

 tubercles would not suffice, we find them transformed into actual 

 hooks ; w hile in some of the succincti they are little more than 

 a thickening of the anterior margin of the sub-joint. In all 

 lepidopterous pupte these remnants of the anal pro-legs are more 

 or less indicated, while in certain moths (Pterophorina:) where 

 the pupa is partly suspended, they are, as in tha Nymphalii/ic, 

 covered with long hooks similar to those at the lip of the 

 cremaster. 



Thirdly, we have what may be called the sustentor ridges, 

 usually connected with the sustainers, embracing them on the 

 outside, and extending backward to the inside of the ventral cre- 

 mastral ridges, and sometimes, as in Paphia ai/d Limenitis, 

 forming quite a deep notch, which doubtless assists in catching 

 hold of the larval skin in the efforts to attach the cremaster. 



^ It is an interesting fact in this connection that Roesel, who has never 

 had any superior as a delineator of insect larva, makes the Nymphalids in 

 his figures all suspend to an elongate conical piece uf silk apparently issuing 

 from the anus, with the legs invariably free and in no instance h^- kcd. It 

 is evident, however, from his text, that he was not aw ..re uf the use of the 

 anal plate, and since he speaks of the larvse atcaching themselves by the 

 hind legs or extremities, it is equally evident that his figures do nol correspond 

 with the text, while the freedom of the legs in his figures is, of c <urse, an 

 error 



2 What is here termed the coriiim is the membran us layer between the 

 separating larva skin and the forming chrysalis. If, as recent inve-ligalions 

 seem to show, it is only the outer half of the derma! layer of the skin which 

 is cast off in the exuviation of invertebrates, and nut the "hole skin with 

 its three layers, then this membrane is developed between the splittiHi; parts 

 of said outer layer, and is not, strictly speaking, the cerium. 



