258 Experimental Zoology 
their early development until the embryo hatches and begins to 
feed for itself. It appears, however, in the case of some water 
animals that the egg or the early embryo derives some of the 
materials necessary for development from the surrounding me- 
dium. In the development of the eggs of the lower marine 
animals it has been shown that certain inorganic constituents of 
the sea water enter directly into the embryo and take part in its 
growth. The most thorough study of this sort that has been 
carried out is that of Herbst. His final conclusion is that for 
normal growth all of the common constituents of sea water are 
necessary. For a few of them other closely related chemical 
compounds may be substituted, but it is surprising how little 
substitution of this sort is possible without altering the normal 
processes of growth. In such cases it must not be overlooked 
that the egg itself is a storehouse of food materials, and that the 
growth consists mainly in adding water to the materials already 
present; but there can be no doubt that extraneous salts are 
also fixed by the tissues of these organisms and take a part 
in the growth. For example, the young embryo of the sea 
urchin produces a calcareous skeleton. Unless calcareous salts 
are in the water, the skeleton is not formed; even other closely 
related salts cannot be substituted. Quite recently Maas has 
carried out an experiment with sponges. He finds in them also 
that unless the calcareous constituents of the sea water are pres- 
ent, no spicules are formed, and the development does not extend 
beyond a certain point. 
By adding salt to fresh water the density may be increased. 
By diluting sea water its density may be decreased. Both 
changes may affect the rate of growth, as shown by the following 
experiments. Yung (1885) added 2.0, 4.0, 6.0, 8.0 grams of sea 
salt to 1000 grams of fresh water, and placed frog embryos in 
the solutions. In the .2 per cent solution the rate was nearly 
the same as in pure water. Retarded development occurred in 
all the others. In the .8 per cent solutions the tadpoles hatched 
17 days after the normal time. 
It is not altogether clear in this case whether the effects are 
