534 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1914. 
found a large empodium and two strong curved claws. The entire 
regenerated part is about 8 millimeters long, of which the tarsal 
part measures 3.7 mm. The normal tarsus in an imago measures 
with its five joints about 6 mm. 
Of the other specimens within this subgroup, two were nearly full 
grown and one was half grown. 
The other subgroup contains only two regenerations, both found in 
full-grown specimens. They differ from the other examples within 
this group mainly by the absence of the tibia-like segment. They 
consist only of four well-developed tarsal joints, which in size are 
equal to the corresponding joints in the cases described above. 
It is possible that these regenerations under favorable conditions 
might develop further and acquire the tibia-like segment, which alone 
differentiates them from the other cases in this group. 
It should be mentioned in this connection that the antennal tarsi, 
like the regenerated tarsi on the legs, at most consist of four joints, 
while the normal tarsus is five-jointed. 
Group C.—This group is not represented among the material pro- 
duced in these experiments, but contains only the spontaneous regen- 
eration, described in the foregoing. This case is sharply differentiated 
from the types just described by the possession of a stem of antenna- 
like joints. Nowhere else in my material have I found any case in 
which I have been able to detect the presence of a stem of antennal 
joints bearing the tarsal joints. 
Another division of the material into two groups may be made— 
for example, on the different composition of the regenerated append- 
ages; the first group containing such specimens which consist only 
of tarsal joints and the other including such which possess a tibia-like 
joint in addition. 
Such a division will not coincide with the first division made on the 
different places of amputation, on account of the grouping within the 
group B, above defined. This may be merely a result of the manner 
in which the amputations were made and of insufficient material. 
It has already been noted that further experimentation may prove 
that the two subdivisions within group B may represent the same 
type in different stage of development. 
With regard to a possible relation between the structure of the 
regeneration and the place of amputation, I shall confine myself to 
giving the general impression I received during my studies: Amputa- 
tions across the basal joint or between this and the second joint produce 
dwarfed, slightly developed homeeotic regenerations, while amputa- 
tions across the second joint or between the second and the third 
joints cause strong well-developed homeeotic regenerations. It is 
possible that quite different results may be reached by more careful 
experimental studies of larger material, which shall obviate all such 
