On the Rhizocephalan Genus Thompsonia, etc. 15 
host from moulting. But some time ago I pointed out (7) that hermit 
crabs infected by Peltogaster still moult frequently and now it is to be 
recorded that, although the root system of Thompsonia is almost as 
extensive as that of Sacculina and the great number of external sacs 
must require a large and constant supply of nourishment, yet no check 
appears to be placed on the moulting function of the host. The 
number of moults observed was, it is true, only two or three, but this 
was due to the difficulty experienced in keeping Synalpheus in captivity 
and to the short time of observation, and I do not doubt that the 
moulting of infected specimens is a normal and frequent phenomenon. 
Not only, then, is the whole number of existing external sacs budded 
off from one root system, but provision is made for replacing these when 
they have functioned. A necessary condition for securing this replace- 
Fig. 6. 
A. Roots of central system, showing external layer, e.l.; with nuclei, n.; lacunar 
space, lac.; yolk globules, y. 660. 
B. Root of peripheral system running through a pleuron of Synalpheus. This 
shows the non-nucleated external layer, the lacunar tissue, lac. tiss., and 
a developing egg, ov. 660. Fixed in corrosive sublimate and stained 
with borax-carmine. 
ment is that the host should continue to moult. Further consideration 
will be given to the whole phenomenon after the sections dealing with 
the external sacs and the development of the egg cells. At present I 
wish to point out that the root system has the remarkable power of 
regenerating external sacs and producing germ cells throughout the 
whole period of its existence. 
If the distribution of the external sacs (which must follow that of the 
root system) is studied in a series of individuals it is seen to vary con- 
siderably. In some hosts they are confined to the abdominal swim- 
merets, but in the majority they are also present anteriorly on the thor- 
acic limbs and posteriorly on the tail-fans and telson. In the former 
lightly infected individuals the root system has, we may suppose, been 
established for a comparatively short period, while in the latter class 
