A STUDY OF METAMERISM. 435 



The results that have, on the whole^ been the most puzzling 

 to me, and which I cannot pretend to entirely explain, are 

 some of those cases where the internal and the external 

 spirals do not agree. It seems quite certain that in the far 

 larger number of cases the false unions are at bottom, the 

 result of imperfect joining of the mesodermic blocks, and that 

 the ectodermic grooves mould themselves on the internal 

 arrangement. 



But in cases of disagreement it seems clear that the outer 

 grooves have to a certain extent an independent individuality, 

 and may behave differently from the spirals in the mesoderm. 

 When we find similar spirals in antennse, &c., it is not, after 

 all, so surprising that the outer body-wall of the worm should 

 show this independent variation. 



IX. Modifications in Antennae of Arthropods. 



The antennae of Arthropods are made up of a series of seg- 

 ments or rings. In some Arthropods the segments are long 

 and relatively few in number. In others the segments are 

 narrow rings and very numerous. The latter kind are more 

 likely to show variations in the arrangement. The antennse 

 and antennules of the lobster (Homarus vulgaris) I found 

 showed many variations. Perhaps the lobster^s antennae and 

 antennules may owe many of these false arrangements to the 

 fact that if they have been lost new ones will regenerate, but 

 that all the abnormalities come from this source is extremely 

 improbable. 



I have found in the antennae of the lobster nearly all the sorts 

 of variation in the arrangement of the rings that are to be 

 found in the earthworm. A few of these variations are shown 

 in PI. 43, figs. 90 to 95. Fig. 90, a, b, shows a compound ring. 

 Fig. 91, A,B, shows a double compound ring, where three half- 

 rings of one side are united to one of the other. Both of these 

 modifications are common, particularly the former, at the broad 

 base of the antennae. 



In nearly all cases the '^splits" run in from the sides. 

 Rarely a modification similar to that shown in fig. 92, a, b, is 



