THE CODIUMS OK THE JUAN l" liKN ANDICZ ISLANDS 595 



made up of distinctive microspecies in the general coiisinship of Dccoriicata. 

 The type group is in the east Atlantic, from whicli region have been proposed 

 Codiitiii di-corticatinn (Woodw.) M. A. Howe and C. Cliazaliei VV. v. B. In the 

 west Atlantic is another species (or possibly group of s[)ecies) not clearly 

 segregated as yet from other species and no distinctive name, or names, have 

 thus far been applied. In the northeast Pacific, several seemingly specific 

 variants appear, such as those designated as C. dfcorhcatitJii {\cvu\r\i), C. aiiip/i- 

 vesiculatuin Setch. et Gard., and C. iinilatfralc Setclu et Gard. In the north- 

 west Pacific occurs C. cylindi icuni Holmes, a species usually more or less con- 

 fused with C. divaricattuti Holmes (of clo)igat2im type). On the southeast 

 Australian and the southeast African coasts are other species which have been 

 confused earlier with C. tomcutosiim (Auctt.) or later with C. clongatnm (Auctt.) 

 and now among the Juan Fernandez specimens, C. foiiaudcryianum is still 

 another member (or group) of Dccoyticata. 



The type of Codiuiii fcrnaiidcziamiui (Setchell, in Herb. Kew, No. 65!), 

 collected by H. N. MoSELEY of the Challenger Expedition in November, 1875, 

 and determined originally as C. tovicutosiiin, is a plant (pi. 36, fig. 8) up to 

 »3 feet» long, rather stout, lacking a base and with a single, seeming uneciual 

 dichotomy. It is cylindrical throughout, with possibly a slight (.') infra-axillary 

 dilation at the base of the only dichotomy present. The nearest approach to 

 it is Skottsberg s No. i. Both the type and No. i might readily be referred 

 to C. decortkatiiin (Woodw.) M. A. Howe (verum!) so far as habit and general 

 utricular structure (see plates 35 and 36, figs. 5 and 8, and plate 42) are 

 concerned, but the »adult» utricles are apparently always much shorter, not 

 exceeding two-thirds of the length characteristic of the »adult» utricles of C, 

 dccorticatiim (verum!). In the broader (euryterophyse) utricles, in this group 

 generally scattered among the narrower »adult» utricles, there are found in 

 C. fernandezianuiii decidedly broader utricles than have been found in 6. dc- 

 corticalnut (verum!). Otherwise than in extremes of their particular dimensions, 

 the utricles of C. feniandeziamim are those characteristic of the Decorticata\ 

 large or megaphyse, usually with one or more verticils of hairs (or hair scars) 

 just below the usually rounded apex and a whorl (irregular) of gametangia 

 between one-third and one-half below the apex. The apical membrane is 

 usually more or less thickened in the utricles of members of the Decorticata. 



In explanation of the number of figures, both of habit and of utricle, of 

 CodiuDi fcniandeziauuni, which it has seemed desirable to reproduce, it may 

 be said that they represent fairly completely the basis for the foregoing con- 

 clusion and will help to explain the difficulties of not only this, but of other 

 Cod'uim problems presented in any attempt to classify the species. A figure 

 of the type (pi. 3, fig. 8) shows the habit of a possibly considerable fragment, 

 with its elongated sparse dichotomies. The figure of SKOTTSBERG No. i (pi. 

 35, fig, 5) shows a characteristic fragment; No. 4 (pi. l6, figs. 9, 10) shows a 

 slender bit (right) and a fragment seemingly having suffered injury and proli- 

 ferated into a broad dichotomy; No. 3 (pi. 37, fig. n) shows a broad fragment, 

 flattened and dilated, possibly also having suffered mutilation; No. 7 (pi. 2, 

 fig. 6) shows a characteristic fragment (A) and one (pi. 35, fig. 7) (K) which 

 under some stimulation has become covered with short, crow^ded proliferations; 



