132 



NATURE 



[June ii, 1891 



from the connecting ducts, and that intermediate condi- 

 tions can be found in which the bars can neither be said 

 to be absent nor present. He describes this condition 

 in his new species Perophora bnnyulensis, and it is also 

 present in P. viridis, Verrill, and in various other Simple 

 Ascidians, as has been shown in the Challenger Reports 

 and elsewhere. 



Such cases, although rather perplexing to the syste- 

 matist, are perfectly natural from an evolutionist's 

 point of view, and they certainly make one regard with 

 some suspicion large groups founded upon any such 

 one character. Consequently, Lahille's order " Stolido- 

 branchiata," characterized solely by the presence of 

 a particular kind of internal longitudinal bar in the 

 branchial sac, is, in my opinion, a most unnatural 

 assemblage of the families Polystyelidse, Cynthiidae, 

 Molgulidag, and Botryllidse, which cannot be retained. 

 It is not safe to trust to the modifications of structure 

 of one organ in the detailed classification of a group, and 

 it is especially unsafe where that organ is, as in the case 

 of the branchial sac, of great physiological importance, 

 and so is liable to be considerably modified in accordance 

 with the mode of life in forms which are otherwise closely 

 related. Morphological characters of less functional im- 

 portance are more likely to be retained unaltered, and so 

 indicate real genetic affinity. 



Surely Lahille does not seriously mean to contend that 

 the internal longitudinal bars in the branchial sac of the 

 Botryllidae, Cynthiidae, &c., are different in any morpho- 

 logical sense from the similar bars found in other Asci- 

 dians, such as the Ascidiidae. Although they may be 

 slightly different^ in their relations to the wall of the sac 

 in these two groups, being attached throughout their 

 length in Botryllits in place of only at the angles of the 

 meshes as in Ascidza, and are therefore somewhat different 

 in their development (ontogeny), there can scarcely be any 

 doubt that in their origin (phylogeny) all such bars in 

 the branchial sac are alike, and are therefore homologous 

 structures. 



(3) It follows from what has been said above in regard 

 to the origin of the Compound Ascidians, that even though 

 the group Polystyelidse is placed (as was the case in the 

 Challenger Report) in the Ascidias Composita;, it is not 

 thereby widely separated from its relations amongst the 

 Simple Ascidians. If the sub-order Ascidiae Compositse 

 is retained, then the PolystyelidcC must go in it, since 

 they form definite permanent colonies with the ascidio- 

 zooids embedded in a common test ; but of course these 

 forms are very similar in many respects to Styela and 

 Polycarpa — that being one of the points of contact be- 

 tween Compound and Simple Ascidians — and therefore I 

 can agree fully with all that Lacaze-Duthiers and Delage 

 say in favour of that relationship. The matter stands 

 simply thus : — If Ascidiae Compositas is retained, the 

 Polystyelidae must be placed in it at the nearest point to 

 Polycarpa 2iVi\on<g'sX Ascidiae Simplices ; while if Ascidiae 

 Compositae is abolished, the Polystyelidae will form a 

 family or a sub-family (it matters little which) alongside 

 the Styelinae under Ascidias Simplices. To go further, 

 and break up even the genera of the Polystyelidae, placing 

 the species beside those Cynthiidae they resemble most in 

 the structure of the branchial sac, would be to give no 

 value at all to the property of reproduction by gemmation 

 and the formation of colonies. 



(4) It has long been recognized that there are two 

 groups of forms in the family Cynthiidae, those which 

 centre around Styela and those related to Cynihid, and 

 when the remarkable stalked forms, such as Boltenia and 

 the deep-sea genus Culeolus, had been added, I defined 

 these three groups as sub-families under the names 

 Styelinffi, Cynthinae, and Bolteninas. Leaving the last 



' Even this difference is not constant. In some Botryllidae, aud I think 

 in all Polystyelidae and many Cynthiidse, the relations of the bars in the 

 adult are precisely as in Ascidia, Ciona. and Ecteinascidia. 



out of the question, we have the two former distinguished 

 amongst other characters by the fact that the Styelinae 

 have never more than eight folds in the branchial sac, 

 and have simple tentacles, while the Cynthinae have always 

 more than eight folds, and compound tentacles. 



A few years ago these seemed well-established characters 

 to which there were no exceptions. Last year, however, 

 Lacaze-Duthiers and Delage published a preliminary ac- 

 count of a Cynthia from the French coasts, with only eight 

 folds (as in Styelinae) in its branchial sac ; while Traustedt 

 has discovered that the Cynthia tesselata of Forbes has 

 four folds on the right side of the branchial sac and three 

 on the left (like some Styelinae), although the tentacles are 

 compound (as in Cynthinae) ; and I find that long ago 

 Alder described the reverse case in Cynthia tuberosa, 

 Macg., where there are twelve folds in the branchial sac 

 (Cynthinae), although the tentacles are simple (Styelinae). 

 Thus the two links required to unite the characters ^ of 

 Styelinae and Cynthinae have been found, which is perfectly 

 natural and satisfactory to the evolutionist, and the 

 question for the systematist now is. Must these two sub- 

 families be united ? I think not. I believe that they are 

 natural groups, and that they are really as widely separated 

 from one another in their typical members as we ever 

 supposed them to be, although not so completely isolated 

 from one another by the extinction of intermediate 

 forms. 



I f these interesting links, to which attention has just been 

 drawn, and which are apparently not common nor widely 

 distributed forms, had become extinct a fev^ years ago, 

 the Styelinae and Cynthinae would without question be 

 justly regarded as widely separated groups. And the 

 present position is merely that a few forms are known 

 which if not bridging over at least lie as stepping-stones 

 in the gap ; while the vast majority of the species in 

 question are clearly distinguishable by easily recognized 

 characters into two definite sets. This last fact has an 

 importance which entitles it to recognition. I am far 

 from wishing to ignore the importance of such inter- 

 mediate forms ; in fact I am more likely, I fancy, to regard 

 them with undue interest ; but after all they are single 

 species, minute twigs of the great branch under con- 

 sideration, while lorg series of typical Styelinae and 

 Cynthinae — the many species of Styela and of Polycarpa, 

 of Cynthia and of Microcosmus — can be divided into two 

 groups by their tentacles and their branchial folds, and 

 I believe we are justified in giving expression to this 

 natural grouping by retaining the two sub-families in our 

 system of classification. It need not lead to any diffi- 

 culties : the intermediate forms can be placed as an 

 appendage to the sub-family taken first. We cannot 

 now pretend to draw hard and fast lines round all our 

 groups, a serial or a tabular classification will always give 

 erroneous impressions, and in a phylogenetic arrangement 

 the linking forms will appear in their proper places as 

 little twigs between the two great branches. 



(5) The genus Botryllus seems to contain an endless 

 series of forms which might be (and many of which have 

 been) described as separate species. Giard, twenty years 

 ago, pointed out the great variability of the species in 

 this genus, and described many varieties and local con- 

 ditions, but the supply is not yet exhausted, and one is 

 almost tempted to conclude that no satisfactory position 

 can be taken up anywhere between the two extremes of 

 either (i) regarding the whole genus (or even the family 

 Botryllidae) as an enormous protean species, or (2) de- 

 scribing nearly every colony as a separate species. 



From the point of view of the systematist or specio- 

 grapher who wants "good" and well-defined species, 

 this group of Ascidians must be an abomination, but 

 to the student of evolution it is full of interest. Here, 

 if anywhere, characters can be seen varying in all 



I ' These are ihe chief characters, tut there are orhers, ?uch as the condition 

 y of the stomach and digestive glands. 



NO. I 128, VOL. 44] 



