340 
Observations on Termites 
dimensions measured are of a different nature, the wasps and termites agree in 
the sexual forms being the larger. 
The absolute variation in the wasps is greater for the workers than for the 
drones and queens. In the case of the five species of termites, the workers have 
a mean absolute variation of 4 67 units against 3'51 for the soldiers and 3'57 
for the male imagos ; that is, a mean of 4*09 for the asexual castes against 3"57 
for the sexual form. The general tendency is consequently for the termites and 
wasps to resemble each other in this matter. 
In the relative variabilities (coefficients of variation) the sexual forms (drones 
and queens) of the wasps are considerably less variable than the workers, the 
means being: workers, 3"55 ; drones, 2'60 ; queens, l"o7. 
In the termites the means are, as we have already seen, for : large soldiers, 
3"26 ; small soldiers, 3'68 ; large workers, 3'74 ; small workers, 3'89 ; male imagos, 
2 06 ; female imagos, 1'99. 
There is thus a ver}' striking similarity in these results : the mean of the 
coefficients for drone and queen wasps is 2"08, and for male and female imagos 
of the termites 2-02. 
The asexual wasp caste of workers with a coefficient of 3"55 also compares very 
closely with the mean of the coefficients of the asexual termite castes, 3"64. 
The material at the disposal of the authors of the wasp paper did not permit 
of the examination of the correlation of the different castes in a number of nests, 
and consequently a comparison on this point with che termites cannot be 
instituted *. 
(13) Sumviary of some of the Results. 
(1) Although the young appear to hatch all alike and in certain species 
{T. natalensis, for example) all are the offspring of one queen and king, yet the 
various asexual and sexual castes of a nest exhibit marked differences in their 
variabilities. The diffei-ences in the variability cannot therefore be regarded as 
due to inheritance, but must be supposed to arise mainly through post-embryonic 
environmental influences. 
(2) The sexual caste is much less variable than the asexual castes. 
(3) The relative variability (measured by the coefficient of variation) of the 
population compared with that of a family (the individuals of a nest) appears to 
be considerably greater than can be accounted for from the ordinary effects of 
inheritance ■}-, and the cause is almost certainly to be found in the moulding influence 
of a varying environment on an exceedingly plastic organism. 
[* The relative variability of the species and of the members of a nest is now being investigated 
and the results will shortly be published. Ed.] 
[t The reduction in variability for the case of assortative mating of a large amount has not yet been 
worked out. For the termites we have in each generation a brother-sister marriage, or an assortative 
mating, say of '5. There is thus only one pair of ancestors in each generation, and the reduction 
of variability for such a system may quite conceivably be as great as that indicated by the termite 
measurements, i.e. nest variability 7'37 and general population variability 17'08. Ed.] 
