COOKE: ON THE GENUS PURPURA. 325 



Page 164. — p. mancinella L. '■ P. aegrota Reeve... is one 

 of these intermediate forms' [between maficinella and ecJiinatci\. 



Tliis is another guess from the look of the figure, in which 

 a touch of yellow at the lips might suggest a reminiscence of 

 mancinella. Aegrota, however, when the type is consulted, 

 turns out to be a very remarkable variety of textiliosa Lam., in 

 which the spire is somewhat produced, and blunt tubercles are 

 formed on the shoulder of the body whorl, while the yellow 

 colour almosts always present on the lip of fextiliosa is slightly 

 deepened. Although the perfect form of aegrota is exceedingly 

 rare, I have noticed specimens of textiliosa with incipient 

 tubercles, which form a passage to that form. 



I am at a loss to understand Vi'hence Tryon has derived 

 his locality, ' W. Coast Central America ' and ' Peru ' for man- 

 cinella. It inhabits the coasts of the Indian Ocean from Natal 

 and Aden to N. Australia and S. Polynesia, but never occurs 

 on the American coasts. 



Page 166. — P. mstica Lam. ''P. niarniorata Pease appears 

 to be the same species.' 



This again looks as if Tryon could not have seen speci- 

 mens of marniorata. What that species really is I do not pre- 

 tend to be able to determine. Not more than a dozen specimens 

 appear ever to have been found (they were dead shells collected 

 by Mr. Andrew Garrett), and those that I have seen are in such 

 wretched condition from the scraping and oiling process which 

 they have undergone that it is impossible to pronounce upon 

 them, except that they bear no affinity to mstica Lam. 



Page 167. — P. haeniastonia Iv. var. iindata Lam. 'By 

 common consent the usug.1 W. Indian manifestation of P. hae- 

 niastovia bears this name, although the form which it character- 

 izes is equally common on the W. Coast of America.' 



Into the vexed question of the identification of the La- 

 marckian P. undata I do not propose to enter, although I suspect 

 that ' common consent ' is here at fault. But it is incorrect to 



