MUSEUM OF COMPARATIVE ZOOLOGY. 261 
even of the same cluster, mature at different times. We must not sup- 
pose, in studying the development of Aga/ma eggs, that the ova found 
free in the water were all cast at the same time. The only trustworthy 
method of observation is to trace individual eggs into larve and time 
their development, which is a most difficult thing to accomplish success- 
fully with these tender creatures. Even if we follow and time with 
care the rate of growth in our glasses, it is a question whether we should 
not make an allowance for retardation or acceleration of this time brought 
about by changes in the temperature of the water in which they are 
placed in our aquaria. 
The disk formed at one pole of the egg by the epiblast, hypoblast, 
and an intermediate transparent layer, may be called the primitive hy- 
drophyllium or covering-scale, to distinguish it from others which are 
later formed. In a stage following the last this body has assumed an 
elevation upon the surface of the egg greater than formerly. As far as 
its general outlines go, no great change has taken place in the larva with 
advancing age; but near the rim of the disk another minute elevation 
in the walls of the epiblast has pushed itself up, which is destined later 
to play an important part in the structure of the adult. This elevation 
(pn. cy.), which at this time cannot be distinguished from a simple bud 
such as any other organ of the Agalma body at first has, is the begin- 
ning of the future float. It is a true bud, as already pointed out by 
Metschnikoff. 
The accounts which Metschnikoff and Haeckel give of the origin of the 
float in genera so nearly related as Agalma and Crystallodes are radically 
different. In Crystallodes, according to Haeckel, the air-sac originates 
from the primitive cavity as a bud. Speaking of changes on the sixth 
day, he says:* “ Die wichtigste Verinderung aber, welche am sechsten 
Tage eintritt, ist die vollstindige Abschniirung des Luftsackes von dem 
Centralraum der Primitivhéhle. Das Entoderm, welches die Wand des 
Luftsackes bildet, und welches bisher an seiner Einmiindung in die 
Centralhshle unmittelbar tiberging nach oben in das Entoderm des 
Deckstiick-Nahreanals, nach vorn in das Entoderm des Polypiten, 
wichset nun vollstiindig an dieser Stelle zusammen. Der Larvenkor- 
per enthilt also nunmehr zwei vollstindig getrennte und geschlossene, 
mit Fliissigkeit erfiillte Hohlen: die einfach rundliche oder langliche 
runde Luftsackhéhle, und die Centralhéhle, welche in vier Caniile sich 
verzweigt, in die Caniile der beiden Knospen, des Deckstiicks und des 
Polypiten. Das Entoderm, welches alle diese Héhlraume auskleidet, 
* Op. cit., p. 58. 
