348 
Races of Entamoeba histolytica 
sistent, we can only suppose that the authors are mistaken in their 
assertion; and that, in the absence of any indication to the contrary, we 
are justified in concluding that the lightness with which they dismiss 
this difficult problem reflects the spirit in which they effected its solution. 
It seems clear, from their published statements, that Mathis and 
Mercier have, like Kuenen and Swellengrebel, ignored all strains of 
E. histolytica save the commonest—those which, as in our Case E. 42, 
produce cysts having a mean diameter between 12 p, and 14p,. The 
smaller and larger strains they have, apparently, completely overlooked. 
The curve which Mathis and Mercier have published (1917 a, p. 168), 
and upon which the proof of a dimorphism in the cysts of E. histolytica 
appears to rest, is to us wholly incomprehensible. As we have seen, the 
measurements which Ave have made of single strains of E. histolytica 
have consistently given us unimodal curves. There has always been a 
common diameter for the majority of the cysts, with the minority 
grouped more or less symmetrically above and below it. Even by taking 
so few as a hundred cysts we cannot understand how it has been possible 
to obtain two modes separated from one another by only T5/x, with only 
half-a-dozen cysts of intermediate dimensions. In no instance have a 
hundred cysts taken at random from our series of measurements given 
us a curve of this form. We were at first inclined to suppose that the 
dimorphism was similar to that which we ourselves have found in Case 
E. 130—that is to say, it resulted from an infection with two different 
strains. This explanation, however, will not suffice: for the superposition 
of the curves from two different strains having their maxima at 10p, and 
lT5p. would not give a curve like that of Mathis and Mercier. It would 
be necessary to assume in addition that the strain having the larger cysts 
showed practically no variation in size below the mean, while the smaller 
showed none above. Another possible explanation is that which we have 
suggested {sufra, Part I, A) to account for the anomalous bimodal curve 
which we obtained from one of our oAvn series of measurements. 
Whatever the correct interpretation may be, it is clear from our own 
measurements that there are many strains of E. histolytica whose cysts 
display no discoverable dimorphism in respect of size. Indeed, it appears 
to us that when a dimorphism of the cysts in any given stool is demon¬ 
strable, it is to be interpreted as the result of a simultaneous infection 
of the patient with two different strains. This very obvious inter¬ 
pretation is not even considered by Mathis and Mercier, though no proof 
that there is a dimorphism of the cysts in any—still less in every—strain 
can be furnished until this possibility has been excluded. For our OAvn 
