8o 



cnidocysts. The lateral filaments are short, horn-shaped and always curled up. They are small 

 compared with the median lobe which has twice the length of the lateral filaments and is 

 quill-shaped. They occur in great numbers at the base of the .siphon; occasionally we found 

 an extended tentacle bearing a great number of these tentilla. 



The tentilla which mcjst resemble our type II are those which H.\eck:kl 88b figures for 

 his Crystallodes vitrea. The size is also very much the same. We found however in some ot 

 our young specimens of group II tentilla which w^ere smaller, whilst in older specimens we found 

 also larger tentilla. Whilst examining this second group, we supposed that in the well-preserved 

 smallest one the siphon which was lying farthest, should be considered the first developed and 

 therefore the oldest. 



We found in Cat. 25 R. two siphons, which bore the same kind of tentilla. A specimen 

 with only one siphon with tentilla did not — to our regret — come under observation. We 

 were very much astonished to find only one kind of tentilla, especially as the specimen was 

 comparatively small. It was only natural to suppo.se that these tentilla type II would repre.sent 

 the definite ones of the group. And we found the following .series of development: Cat. 25 R., 

 175 G., 122 A., 176 C, 215 D. showed two developed siphons with tentilla type II, Cat. 25 O.; 

 88 J., 176 B., 24 B., 25 N., 176 D., '93 A., 215 O., 24 C, three mature siphons with the same. 

 Cat. 88 H. and 176F. as many as four siphons with tentilla type II. 



To our astonishment we found that the following Catalogue-number 151 A. consists of 

 four siphons with tentilla type II and a fifth with young tentilla type IV, identical 

 with the same in the first group of Crvstalhnnia. 



It seems difficult to believe that all these four siphons with tentilla type II are provisionally 

 ones, especially as the third siphon in Cat. 151 A. already shows the development near its base 

 of a young female gonodendron. We should then have to consider those specimens in group I 

 which have lost their primary siphon in the .same stage of development as those in group II 

 which have lost all their 4 siphons; in our series it would be the very last Cat. 33 A. 



A larger collection of complete material may perhaps give a less complicated view of 

 the question. 



At any rate we tried to continue our .series as we had done in group I, putting under 

 151 A. those specimens which showed less than four siphons with tentilla type II and so on, 

 up to 33 A. which has none left. 



A .stage of abnormal development of either of the oldest sijihon or its tentilla was 

 unfortunately not represented. This classification of our specimens of course is not a definite one. 



The different specimens of this second group show more variation in shape from one 

 another than in the first group. A glance at the 6 specimens of Crystallomia spec, group II 

 figured on PI. XIV, figg. 106 — 109 and PI. XV, figg. i 10 — 112 shows clearly the accuracy of 

 this statement. 



Cat. rj§ G. PI. XIV, fig. 106 (25 R. is not sketched as its appendages were very much 

 altered b\- the preserving fluids) has a nectosome of I'^j, mm. length, a siphosome of about 

 2 mm. The pneumatophore is small, elongate, with a little purple pigment cap at its top. It 

 has not the beautiful shape which we find in the young sjjecimens (f i. Cat. 51 A.) of group I. 



