630 ■ Transactions. — Geology. 



species ; but the present large collection — all from one place — 

 enables me to pursue another system, and to ascertain which 

 bones belong to different species by the method of averages. 

 The advantage is that we thus find the commonest or most 

 typical form of each species, while by the former method we 

 may be taking an exceptional bird as our guide. 



The result of my examination is to show that, although 

 the species do, undoubtedly, often pass into each other, still 

 the connecting links are comparatively rare, while the main 

 body of the individuals of a group are well separated from the 

 main body of individuals of another group ; or, in other words, 

 the individuals form clusters with only a few connecting links. 

 These clusters or groups I take to be species ; and this has 

 made me abandon the following specific names formerly pro- 

 posed by me, although some of them may, perhaps, have to 

 be used in the future as varietal names : Dinorms validus, 

 potens, and strenuus ; Palajjteryx i^lena. ; Euryapteryx com- 

 pacta ; and Pachyornis valgus. 



The classification of the moas has been further compli- 

 cated by the idea that the sexes must have been of different 

 sizes. This idea, discussed and rejected by Sir Eichard Owen,* 

 was again brought forward by Sir Julius von Haast, and was 

 subsequently developed by Mr. R. Lydekker. It has, how- 

 ever, no positive evidence in its favour, but, as Mr. Lydekker 

 himself says, is an inference from the case of the Apterygida, 

 in which the females are larger than the males. But the 

 measurements which I have made of 2,010 leg-bones of the 

 Dinornitldda by no means confirm this inference; on the con- 

 trary, they show distinctly that, in several species at least, 

 sexual differences of size are very slight, if any ; and this 

 result I take to be the most important part of my investiga- 

 tion. The evidence for this statement will be given in the 

 remarks on the species, but it will save much repetition if I 

 here give an outline of the reasoning employed. 



In the first place, if the sexes of a species are of different 

 sizes, these two sizes will be more numerous than other sizes 

 which are individual variations from the average, and a 

 species would consist of two groups or clusters. But we must 

 remember that this might also be due to the existence of two 

 varieties of different sizes, too closely connected by inter- 

 mediate sizes to allow us to call them separate species. 



In the second place, we may fairly assume that the sexes 

 lived together and had the same habits. If the moas were 

 monogamous — as are the kiwi, the emu, and the cassowary — 

 we should expect each sex to be equally numerous. If, how- 

 ever, they were polygamous — as are the rhea and ostrich — 



* Trans. Zool. Sjc. of London, vol. iii., p. 252. 



