1898.] • SPECIES OY MILLEPOUA. 253 



It will be seea from these figures that there is not much 

 variation in the average proportion of dactylopores to gastropores 

 in the different forms examined. The largest number of cycles 1 

 was able to count oq one colony gave au average of a trifle under 6. 

 It is noteworthy that this is the exact mean of the highest and 

 lowest averages obtained from smaller specimens on which only a 

 few cycles could be counted. 



The extreme averages 5-08 and 7*08 (IV. & V.) do not show 

 so great a range as may be seen on different parts of a single piece 

 9 and 4, and 8 and 3. 



On the basal incrusting regions of a specimen of Millepore in 

 the Manchester Museum I have observed several widely-separated 

 gastropores attended by only one, two, or three dactylopores, and a 

 similar paucity of dactylopores I have more recently noticed in 

 specimens from the collection made by Mr. Gardiner in Funafuti 

 aud Eotuma. 



I may point to the figures obtained from an examination of the 

 specimens of M. alcicornis given to me by Mr. Lister to show the 

 variability of this feature in the colony. 



The specimens were a number of broken branches, each a few 

 inches in length, beautifully preserved in spirit. Two specimens 

 were taken at random and twelve cycles counted on each. The 

 average of one came out 6-7 dactylozooids to each gastrozooid, and 

 of the other 5-08 dactylozooids to each gastrozooid. 



The only author who has referred to the number of dactylopores 

 in each cycle is Moseley. He says that each group consists " of a 

 centrally placed gastropore surrounded by a ring of five, six, or 

 seven dactylopores," and oq counting the number of dactylopores 

 in each cycle that are drawn in Mr. Wild's picture in Moseley 's 

 ' Philosophical Transactions ' paper I find that the average is 6. 



In Milne-Edwards and Haime's figure of M. intricata there are 

 5 gastropores to 35 dactylopores ; of M. verrucosa, there are 7 

 gastropores to 32 dactylopores (?) ; in M. tuberculosa, 5 gastropores 

 to 18 dactylopores ; but it is not certain that these figures can be 

 absolutely relied upon. They are, however, on the whole, very 

 similar to my own results. 



The general conclusions, then, that must be drawn from these 

 observations are : — 



That the number of dactylopores in each group is very variable 

 in each individual colony of Millepora. There may be, in fact, 

 any number up to 8 or 9. 



That specimens of widely different forms of growth have ap- 

 proximately the same average number of dactylopores in each 

 group. 



That the average number of dactylopores in each group for 

 specimens of all kinds is about 6. 



That the average number of dactylopores to each gastropore 

 cannot be used as a specific character. 



Anatomy of the Soft Parts. — I have examined the anatomy of the 

 soft parts of a large number of specimens preserved in alcohol by 



