ON MGRISIA LYONSI. 371 
life with only four per-radial ones ; in such cases the number 
of tentacles would increase with the age, and therefore with 
the size of the individuals, This certainly is not the case 
with the medusz which I have been describing, for in these 
they are chiefly the smaller individuals which possess super- 
numerary tentacles, the percentage of variability being very 
different in the smaller and larger specimens. 
in order to show this point I divided the medusze which 
I examined into two groups :—(a) With an umbrella- 
diameter varying from 4 to 2 mm., and (b) with a diameter 
measuring from 2} to 4 mm. 
Group (a) contained 278 individuals, of which 39, or about 
14 per cent., had supernumerary tentacles, whereas of the 122 
larger forms belonging to group (b) only 5 possessed such 
structures. 
I think the figures show clearly that the multi-tentacular 
forms must be considered as exceptional varieties ; from such 
small numbers it would not be fair to argue that some sort 
of natural selection is in progress. 
III. Systematic Posrrion. 
The systematic position of Mcerisia lyonsi, unlike that 
of the other known lacustrine medusz, presents no difficulty. 
The globular shape and four-rayed symmetry of the um- 
brella, the manubrial gonads, and the absence of otocysts 
refer the medusa to the Order Anthomeduse. 
The gymnoblastic hydroid stage confirms this position, 
The simple mouth, the four unbranched tentacles, and the 
- narrow radial canals are a combination of characters which 
exclude Moerisia from Haeckel’s (11) familes Tiaride, 
Margelidez, and Cladonemide, and refer it to the only 
remaining family, that of the Codonide, to certain.genera 
of which (e.g. Sarsia) it bears a striking resemblance. 
The arrangement of the gonads in the adult is rather 
different to that of the typical members of the family. The 
study of young individuals, however, removes this difficulty, 
