458 ROSS G. HARRISON 
(table 1). When all scattering mesoderm cells are removed, 
the limb develops in but 36.5 per cent of the cases. 
Covering of the wound with ectoderm from the flank further 
reduces the probability of regeneration. For wounds of 3 so- 
mites in diameter the percentage of positive cases is reduced 
from 52 to 33, and development is entirely prevented by covering 
wounds of 33 somites or over. 
Even when the free appendage fails to develop, parts of the 
shoulder girdle are formed. 
It is concluded that in these operations the cells that normally 
give rise to the limb are removed. In the subsequent process 
of wound healing cells in the surrounding zone move in and 
ultimately form a new limb bud. Around the limb-forming 
cells there is thus a zone of tissue which has the power, in grad- 
ually diminishing intensity toward the periphery, to form a limb 
vicariously. 
When half of the limb bud is removed, disturbances of de- 
velopment may occur, but in many cases normal development of 
the limb follows. 
These disturbances vary all the way from complete suppres- 
sion of development to slight retardation. Abnormalities, such 
as defective digits or hand and reduplications, also occur. * 
There is no distinct correlation between the part of the limb 
bud removed and any particular defect. 
Reduplications are most frequent after extirpation of the 
anterior half. 
Normal development may occur after excision of any half of 
the limb, but,is more frequent in case of removal of the ventral 
half. 
The limb-forming material is divided into approximately 
equal parts by lines anterior and dorsal, respectively, to the 
vertical and horizontal diameters of the limb dise 33 somites in 
diameter (fig. 2). 
On the whole, these experiments show that, while quantita- 
tively the limb-forming tissue is unequally distributed in the four 
quadrants of the area usually extirpated, there is no qualitative 
difference in the potencies of the cells of the four quadrants. 
