398 
POPULAR SCIENCE REVIEW. 
in both the Trilobites and Phyllopods we find genera in which 
a greater number of segments is attained than the normal 
number in the Crustacean type. Thus in Gonocepkalus there 
are sixteen (^reckoning only one each for the head-shield and 
pygidium), in Paradoxides twenty to twenty-two, in Arethusina 
twenty-two, and in Harpes twenty-eight ; and if we look upon 
the head-shield and tail-plate as composed of several body-rings 
united together — which seems certainly to be the correct view — 
we have forms presented to us in which the multiplication of 
segments like each other is one of its peculiar features, illus- 
trating that form of growth which Professor Owen has most 
aptly described as a vegetative repetition of parts. 
Barrande, who has made a special study of this group, when 
writing recently upon the divisions of the body, divides them 
into four groups : 
The first with from 1 to 4 free and movable thoracic segments, 
containing 2 genera. 
The second „ 
5 to 9 „ 
ii 
a 
a 
24 
The third „ 
10 to 13 „ 
a 
a 
a 
32 
The fourth „ 
14 to 26 „ 
a 
a 
a 
16 
We thus perceive that those forms of Trilobites having a 
great excess of free segments is not large when we consider the 
whole as a group. 
In the higher and more specialised forms of Isopoda of the 
present day we do not find the number of segments absolutely 
adhered to without any variation ; on the contrary, we con- 
stantly meet with individuals in which more or fewer segments 
are welded together, so as to conceal the normal number of 
seven thoracic somites between the head and the abdomen. 
Such being the case, we cannot be surprised to find con- 
siderable variation in a group like the Trilobita, which, if they 
really are the remote ancestors of the recent Isopoda, must be, 
according to the views I have suggested above, the prototypes 
of the larvse rather than of the adult stage of the living form. 
Dr. Dohrn and other writers upon the Arthropoda have 
pointed out the remarkable similarity between the larval 
stages discovered by Barrande of certain forms of Trilobites, of 
which we have reproduced two sets, namely Trinucleus or- 
natus , Sternb., and Sao hirsuta , Barr. (See Plate XCI., figs. 
1 - 8 , and 9-15), and the larval stages observed in the living 
Limulus (See Plate XCI., figs. 20-24). But the larvse of 
the Trilobites pass through these stages after being hatched, 
whilst those of Limulus take place in the egg. Bearing in 
mind that in the Trilobites we are dealing with animals pos- 
sessing in many cases a large number of free thoracic segments 
covered by a firm calcareous crust or shell (at least it is so in 
