of Edinburgh, Session 1870-71. 513 
umbella, at the base of the manubrium, to be expanded laterally, 
and the gelatinous extoderm of its floor to be projected along four 
or eight symmetrically disposed radiating lines into as many thick 
pillars, which converge towards the axis, and there meet the manu¬ 
brium, while the thin intervening portions between the pillars 
become developed into generative pouches, the velum at the same 
time disappearing. A hydroid medusa would thus, in all essential 
points, become converted into a discophorous medusa. 
A Lucernaria was conceived of by imagining a Hydra to have its 
tentacles reduced to four in number, and expanded laterally until 
their sides meet and coalesce; while the hypostome continues free, 
the solid hydrorhizal basis becoming at the same time extended 
into a peduncle of attachment traversed longitudinally by four 
canal-like prolongations of the body cavity, or else by a simple 
continuation of this cavity. 
Lastly, a Beroe was taken as a type of the Ctenophora , and was 
conceived of as a hydroid medusa so modified as to become reduced 
to the atrial region alone. The two lateral canals which spring 
from the somatic cavity in Beroe , and subdivide so as to form ulti¬ 
mately the eight meridional canals, correspond to the greatly deve¬ 
loped basal portion of the radiating canals of the medusa, or that 
portion of those canals which is still contained within the solid 
summit of the umbella; the affinities of the Ctenophora being thus 
directly with the Hydrozoa instead of the Actinozoa. 
The author finds the key to the homology of Beroe , and the tran¬ 
sition between the Ctenophora and the Hydroida in the singular 
ambulatory gonophore of Clavatella. 
