730 TRANSACTIONS OF SECTION D. 



been developed some differentiation or segregation must occur by which the two 

 kinds are constituted. If this differentiation should prove to be analogous to the 

 segregation which occurs during the formation of gametes we should be able 

 to account for much that is at present perplexing in the polymorphism of the 

 Cheilostomata. We should in particular not be precluded from regarding a 

 colony with avicularia of one type as nearly related to other colonies which 

 possess avicularia of the other type ; and we should have some explanation of the 

 fact that many of the genera possess all the different forms of avicularia which 

 are variously distributed among their constituent species. 



I have so far spoken as if the adventitious avicularia belonged to two types 

 only. This statement requires some further qualification, although it may never- 

 theless be true that all the forms can be referred to one or other of the two 

 principal kinds. As a matter of fact a single Cheilostome colony may bear more 

 than two sorts of avicularia; as, for instance, appendages with large pointed 

 mandibles, in addition to two kinds of those with small rounded mandibles.* 

 This introduces a further complication, about which it is unnecessary to speculate 

 at present. 



It may naturally be asked whether there are any numerical facts which 

 support the suggestions I have made with regard to the significance of the 

 different forms of avicularian appendages. 1 must admit that the numerical 

 relations are so complicated and apparently so variable that I have not been able 

 to draw any definite conclusion from them. 



Experimental evidence is at present wanting, nor would it be easy to devise 

 crucial tests. Even if it were possible to experiment with two colonies of the 

 same species which differ in their avicularian appendages, the result might be 

 negative, since it is not possible to say definitely whether the eggs of a given 

 colony are normally fertilised by the spermatozoa of the same colony or by those 

 of a different colon)*. Some light may conceivably be obtained from observations 

 on the regenerative processes which may occur in Polyzoa. A recent paper by 

 Levinsen ■^ gives some information with regard to this point, and there are a few 

 other observations on the same subject scattered through the literature of the 

 Polyzoa. 



It is thus obvious that the speculations in which I have permitted myself 

 to indulge cannot be regarded as more than a guess as to the significance of 

 the causes which underlie the facts observed ; but, whether the view I have 

 outlined has anything to recommend it or not, the observations on which I have 

 depended are, I think, correct. If this be the case, some explanation of the facts 

 is urgently required. The decision of the principles on which the Polyzoa should 

 be classified may not be a matter of immediate practical importance, but our 

 theories of species cannot be regarded as established until they have shown them- 

 selves capable of explaining all the cases. Some modification of the Mendelian 

 theory seems to me to be capable of elucidating the apparently haphazard way in 

 which the several forms of avicularia are distributed in the species of Cheilostomata, 

 and it may perhaps be allowed to afford a working hypothesis that can be used in 

 systematic study. The results of such a hypothesis would, I think, be far-reaching. 

 Whether we are justified in accepting it provisionally or not, I am convinced that 

 we require some hypothesis by which we may regard two specimens as belonging 

 to the same species, even though they may differ in what might at first sight seem 



' In the species of Retepora, for instance, there may occur the following types 

 of avicularia, in addition to others that need not be mentioned : Conspicuously 

 large avicularia, some of which are usually fenestral, either pointed (a) or 

 rounded (J) ; small avicularia, either pointed (c) or rounded, these latter occurring 

 as two well-marked types in which the mandible is respectively longer than 

 broad (d) or broader than long (e). The following combinations may occur in 

 individual species or colonies: a + o + d, a + d + e, a alone, b + c, b + d, and others. 

 Examples of some of these combinations may be seen in Busk's Keport on the 

 Polyzoa collected by H.M.S. Challenger (Part XXX., 1884). 



^ ' Sur la Regeneration totale des Bryozoaires,' Acad. Roy. des Sci. de Danemark, 

 Bull, do Vannee 1907, N" 4. 



