TRANSACTIONS OF SECTION D. 835 
permanently in fresh water an animal must either be fixed, or else be strong enough 
to withstand and make headway against the currents of the streams or rivers it 
inhabits, for otherwise it will in the long run be swept out to sea, and this con- 
sideration applies to larval forms equally with adults. 
The majority of marine Invertebrates leave the egg as minute ciliated larve: 
and such larvz are quite incapable of holding their own in currents of any strength. 
Hence, it is only forms which have got rid of the free swimming ciliated larval 
stage, and which leave the egg of considerable size and strength, that can establish 
themselves as fresh-water animals. This is effected most readily by the acquisition 
of food yolk—hence the large size of the eggs of fresh-water animals—and is often 
supplemented, as Sollas has shown, by special protective devices of a most interesting 
nature. For this reason fresh-water forms are not so well adapted as their 
marine allies for the study of ancestral history as revealed in larval or embryonic 
development. 
Before leaving the question of food yolk, reference must he made to the 
proposal of the brothers Sarasin, to regard the yolk cells as forming a distinct 
embryonic layer, the lecithoblast,! distinct from the blastoderm. I do not desire 
to speak dogmatically on a point the full bearings of which are not yet apparent, 
but I venture to think that this suggestion will not commend itself to embryologists. 
The distinction between the yolk granules and the cells in which they are embedded 
is a real and fundamental one; but I see no reason for regarding the yolk cells as 
other than originally functional endoderm cells in which yolk granules have 
accumulated to such an extent that they have in extreme cases become devoted 
solely to the storing of food for the embryo.? 
Of all the causes tending to modify development, tending to obscure or falsify 
the ancestral record, food yolk is the most frequent and the most important; its 
position in the ege determines the mode of segmentation ; and its relative abun- 
dance affects profoundly the entire embryonic history, and decides at what 
particular stage, and of what size and form, the embryo shall hatch. 
The loss of food yolk is another disturbing element, the full influence of which 
is as yet imperfectly understood, but the possibility of which must be always kept 
in mind. It is best known in the case of mammals, where it has led to apparent, 
though very deceptive, simplification of development; and it will probably not be 
until the embryology of the large-yolked monotremes is at length described, that we 
shall fully understand the formation of the germinal layers in the higher placental 
mammals. 
Amongst invertebrates we know but little as yet concerning the effects of loss 
of food yolk. It has been suggested that the extraordinary nature of the segmen- 
tation of the egg of Pertpatus capensis, made known to us through Mr. Sedgwick’s 
admirable researches, may be due to loss of food yolk ; a suggestion which receives 
support from the long duration of uterine development in this case. 
Our knowledge is very imperfect as to the ease with which food yolk may he 
acquired or lost; but until our information is more precise on this point, it seems 
unwise to lay much stress on suggested pedigrees which involve great and frequent 
alternations in the amount of food yolk present. 
Of causes other than food yolk, or only indirectly connected with it, which 
tend to falsify the ancestral history, many are now known, but time will only 
permit me to notice the more important. These are distortion, whether in time or 
space ; sudden or violent metamorphosis ; a series of modifications, due chiefly to 
mechanical causes, and which may be spoken of as developmental conveniences ; 
the important question of variability in development ; and finally the great problem 
of degeneration. 
Concerning distortions in time, all embryologists have noticed the tendency to 
anticipation or precocious development of characters which really belong to a later 
1P. and F. Sarasin, Lrgebnisse naturwissenschaftlicher Forschungen auf Ceylon. 
Bd. ii. Heft iii. 1889. 
2 Cf. E. B, Wilson, ‘The Development of Renilla,’ Phil. Trans. 1883, p. 755. 
