64 The American Naturalist. [January, 
owing to extensive torsions and cell shifting, material is brought from 
one side to the other. 
Again, from these studies it seems evident that in these Bratachia 
there is no such definite cell homology to be traced as in the Annelids. 
The tendency of the paper is decidedly iconoclastic; what new gen- 
eralizations may be built up from such data remains for the future to 
reveal. 
Development of Sponges.—An important but much delayed 
paper by H. V. Wilson has at length been published.’ 
The author studied the marine sponges, Esperella, Tedania and 
other genera, in the Bahamas and on the New England coast. 
The main body of this monograph of one hundred pages and twelve 
plates is an account of the formation and metamorphosis of the “ gem- 
mules” in the above named sponges. 
These gemmules in Esperella are masses of cells that contain yolk 
and may appear anywhere in the mesoderm; they are formed by a col- 
lecting of cells to make a central mass or gemmule surrounded by a 
follicle. As the cells multiply other small gemmules may be added on 
to make a compound mass that ultimately undergoes a process of sep- 
aration into the component cells. A free swimming larva is formed by 
this mass of cells that then escapes from the mother tissue. The outer- 
most cells form an ectodermal layer that is ciliated except at the poste- 
rior pole where the cells are flattened and not pigmented as are the 
ciliated cells. The inner cells are connected by processes and form a 
parenchyma mass. 
This active larva attaches itself obliquely by the posterior pole and 
undergoes a metamorphosis in which the ectoderm changes into a layer 
of simple flat cells and the loose cells of the mesodermal parenchyma 
arrange themselves about inter-cellular spaces, that at first are all alike, 
but subsequently become the various sponge cavities, subdermal spaces, 
afferent and efferent canals, These spaces are at first independent and 
only later become connected. The flaggellated chambers also arise in- 
dependent of one another and of the other spaces and by a similar pro- 
cess of cell arrangement about intercellular spaces. 
These remarkable gemmules closely resemble larvæ derived from 
eggs and have “ germ” layers and the same specialization of the poste- 
rior end as is found in the egg larva. This noteworthy resemblance, 
the author thinks is due to inheritance from a common source and that 
: thus the non-sexual gemmule has retained ancestral traits. No other 
5 Journal of Morphology, ix, Sept., 1894. 
