454 H. J. MULLER 
V. THE MUTATIONS OF W AS QUANTITATIVE FLUCTUATIONS OF 
THE GENE 
It is commonly imagined that genes may vary quantitatively, 
and that mutations often consist merely of such quantitative 
changes, although perhaps in a more extreme form than usual. 
Like most other quantitative changes, these supposed variations 
are often thought of as being themselves subject to a deviation 
caused by ‘chance; if so, the variations, when plotted with 
respect to number and magnitude, must form the well-known 
bell-shaped ‘probability curve.’ Enough mutations have now 
been observed in the gene W to determine whether or not these 
variations may be grouped in such a distribution. 
The three genetically tested mutations reported in this paper, 
together with the seven previous mutations of W that were 
fully tested, make a total of ten. A comparison of the homo- 
zygous stocks of these mutants shows that the series of allelo- 
morphs stands about as follows, arranging them in the order of 
their color from darkest to lightest: 1) red (wild type), 2) 
coral, 3) blood, (4) cherry, (5) ivory, (6) tinge, (7) buff, (8) 
écru, (9 and 10) white, and new-white (11) deficiency ultra- 
white (judged by heterozygotes). The colors of 5 to 8, inclusive, 
are of so’ very nearly the same intensity that their order is 
somewhat uncertain, but they are of not quite the same quality. 
The several second mutations (like the mutations of white which 
produced eosin) are not included in this enumeration, which is 
concerned only with variations from a fixed base. It may be 
added, however, that eosin would come between 4 and 5 in the 
above series, the homozygous eosin female being like the cherry 
(4), the eosin male being lighter than this, but darker than the 
ivory male. 
Of the ten certain variations of W, it is first noticeable that 
all were ‘minus’ variations—i.e., lighter than normal, in spite 
of the fact that ‘plus’ variations in eye color can and do occur 
in other loci. This immediately removes the curve of variation 
from the ordinary symmetrical bell-shaped class, and would 
make it an extreme case in the class of ‘skew curves.’ The sec- 
ond outstanding feature is that three of the ten mutants were of 
