FEOM THE INDIAN AND PACIFIC OCEANS. 19 



papers of Kiikenthal (30), Holm (20), and Burcliavdt (5), I feel convinced the foundation 

 on which Kiikenthal has huilt his three main divisions touches a question of primary 

 importance. Just as an oak or an elm or a beech stripped of its leaves shows a totally 

 different mode of branching, so, I believe, the way in which the polyps are grouped on 

 the terminal twigs, the mode of aggregation of the twigs to the hranches, of the branches 

 to the stem are characters of subgeneric value. On the further details of shape I feel 

 less stress should be laid, for I am of Professor Hickson's opinion that the exact form 

 and mode of branching of a colony is iufluenced to a large extent by the external con- 

 ditions. Since that is a subject on which we can get no positive information, it seems to 

 me safer to neglect details of form and shape and to confine the specific diagnosis to 

 the mode of aggregation of the twigs and branches. For example, while I consider the 

 difference in the form of growth between Spongodes thomsoni and Spongodes chimmoi 

 sufficient to be of subgeneric inrportance, such distinctions as to whether the polyp- 

 bearing surface is a long oval or a broad oval or practically spherical are highly 

 unsatisfactory. I am unable to add anything as regards the anatomical and histological 

 characters, owing to the poor state of preservation of my material. But I believe that a 

 reliable character on which to base specific groups is to be found in the spiculation. 

 The sterile stem bears spicules totally different in character from those of the branches, 

 and the spicules of the inner partition-walls are different again from those of the outer 

 wall. The armature of the polyp-heads of a colony shows a general agreement in the 

 arrangement of the spicules, although there may be a considerable amount of variation 

 in the individual polyps. The variation between the polyps of different colonies is often 

 very striking, as is obvious from a comparison of such forms as S. argeutea., S. claoata, 

 S. hicksoni, or S. kbllikeri, all examples of the division Glomerate. These variations 

 are accompanied by differences in other characters. Taking the spicules of the partition- 

 walls : argeutea has broad warty spindles l - 2x03 mm. in size; clavata has merely little 

 rods 0'06 mm. long ; Mcksoni and Iwllikeri both have spindles much the same size, 

 20x0 35 mm. and 20x02 mm. being the given measurements, but those of hicksoni 

 are mentioned as being hooked and curved. In the branch-walls argeutea has slender 

 thorny spindles; clavata has two kinds — spindles 2-6 mm. long, and numerous small 

 thorny spindles 0'2 to 1-2 mm. in length; in both hicksoni and Iwllikeri the spicules 

 are set transversely, but the difference in size is 28 X 022 mm. and 00x019 mm. 

 respectively. These examples leave little room for doubting that the four species are 

 quite distinct and are rightly separated. It is when such characters as these are not 

 consistently different or consistently similar that the task of species-making becomes so 

 difficult. An added difficulty is the variation between different described specimens of 

 the same species. For example, in the British Museum there are two specimens labelled 

 S. divarkata, in one of which the body of the polyp is unarmed, the tentacles 

 only bearing spicules; in the other each point has one large projecting spicule and two 

 or three smaller ones inclining towards it; in neither case do they answer to Kiikenthal's 

 description of a polyp armed with " Doppelreihen von je fiinf spitz konvergierenden 

 Spiculapaaren." Again, the one with no spicules in the polyp has large spicules in both 

 the outer and inner walls, the other has small spicules in the outer walls and none at all 



3* 



