1. Temperature gradients were passed through the developing frog's egg and embryos. These gradients were applied either (a) apico-basally, when they were either (i) adjuvant, or (ii) antagonistic to the egg's own main gradient; or (b) transversely to the egg's main axis--lateral gradients.
2. (a) By this means considerable modification of segmentation and of cell size was induced, and was especially marked in the mid-blastula. Adjuvant gradients accentuated the normal differences in cell size between the animal and vegetative poles. Antagonistic gradients produced a double gradient in cell size, the smallest cells being in the region of the equator, and animal cells, in extreme cases, larger than yolk cells.
(b) Several cases of the non-formation or obliteration of the blastocoel were obtained by all methods of treatment.
(c) Too high temperature with adjuvant gradient produced inhibition at the animal pole, the large retarded cells being very sharply marked off from the surrounding small cells.
(d) Lateral gradients produced a great difference in cell size on the two sides of the eggy and, as in the cases of "inhibition," a sharp line of demarcation may appear between the large cells of the cooled side and the small cells of the heated side.
(e) When two sets of exactly similar eggs were treated simultaneously in opposite ways, then those subjected to the adjuvant gradient were always, at the close of the experiment, at a more advanced stage of development than those subjected to an antagonistic gradient. Because of this the yolk cells of the "adjuvant" eggs were smaller than those of the "antagonistic" eggs, although the former were cooled and the latter heated.
(f) There seems to be a slight permanent effect of the gradient applied during segmentation. Eggs treated with antagonistic gradient tend to develop into microcephalous tadpoles and vice versa.
3. (a) Antagonistic gradients during gastrulation cause a reduction of the gastrular angle. (For definition see Bellamy (1919).)
(b) Antagonistic gradient causes the eggs to gastrulate sooner than adjuvant eggs under exactly similar experimental conditions.
(c) In the neurula stage the differential effect of the gradient is seen in the inhibition of the head and dorsal region in those subjected to antagonistic gradient, and inhibition of tail and ventral region in those subjected to adjuvant gradient.
(d) Whether this alteration of relative sizes of head and tail regions is maintained in later development has not yet been ascertained.
(e) Eggs exposed to lateral gradients in all stages of gastrulation showed marked asymmetries, some of which were apparently regulated later, while others persisted till the death of the tadpole.
4. Side-to-side treatment in the tail bud stage caused the development of marked asymmetry as the result of differential growth of the two sides. As in the case of 3 (e) some tadpoles appeared to regulate back to normal, whereas others remained markedly asymmetrical till death.