504 EDITH M. PRATT. 



I. INTRODUCTION. 



The genus Sarcophytum was founded by Lesson in 1831, its distinguishing feature being 

 the mushroom-shape of the colony. Klunzinger in 1877 included in the genus a form which 

 is not mushroom-shaped, and which he termed Saixophytum patcciflorum. Marenzeller in 1886 

 separated those forms which are mushroom-shaped from those which are not, and for the 

 reception of the latter he established the genus Lobophytum, of which he described several 

 new species. This genus was based upon the form of the colony, the dimoi-phic character 

 of the zooids, and the form and size of the spicules. 



When examining examples of several species in the collection which have been attributed 

 to Lobophytum, I observed that while all the specimens have the lobed fonn characteristic 

 of the genus, some differ from others in that they are extremely hard and brittle. This 

 I found is due to the presence of enormous spicules which occur throughout the colony in 

 very considerable numbers. In some cases siphonozooids were observed to be very minute, 

 and could only be distinguished with the aid of the microscope, while in a few instances 

 they were apparently absent. 



In addition to the abundant supply of excellently preserved material in the collection 

 from the Maldive Islands, the zoological laboratories of the Owens College contained other 

 collections of similar material, which through the kindness of Professor Hickson were placed 

 at my disposal. These included Mr Stanley Gardiner's collection from Funafuti, Dr Willey's 

 collection from New Guinea, New Britain and Lifu, and Professor Haddon's collection from 

 the Torres Straits. I have also been allowed to make a cursory examination of Professor 

 Herdman's collection from Ceylon. It was, therefore, possible for me to make an investi- 

 gation of the anatomy of the genus Lobophytum, which hitherto had not been attempted, 

 and also to make a comparative study of the anatomy of many species, which have been 

 attributed to this genus, from many different localities. The general results of these in- 

 vestigations are contained in the present paper. A more detailed account of the comparative 

 minute anatomy will, I hope, be published later. 



An interesting and important result of my research was the discovery of a large and 

 extremely well-marked superficial canal system in those species of the genus Lobophytum 

 with minute siphonozooids, which occurs also in those with apparently no siphonozooids. In 

 the former this canal system was observed to be in intimate connection with the minute 

 siphonozooids, but it is absent or only very feebly developed in those species with large 

 siphonozooids. Further investigation led me to the conclusion that the present definition 

 of the genus, which is based upon the form of the colony, dimorphism and the character 

 of the spicules, is artificial and inefficient, because it is founded upon insufficient evidence. 



All the species of the genus Lobophytum divide naturally into two groups ; one includes 

 all those species whose anatomy is similar to the species pauciflorum, which has no special 

 superficial canal system, but is provided with comparatively large siphonozooids ; the other 

 group includes all the remaining species of the genus, whose general anatomy is similar to 

 that of the species densum, a form with minute siphonozooids and a large superficial canal 

 system. The important difi'erences between the two groups of species clearly indicate that 

 they are genetically as distinct from each other as each is distinct from the genus Sarco- 

 phytum or any other genus of the Alcyonaria. It is, therefore, necessary to separate the 



