318 
Merism and Sex in " Spiiiax Niger 
Sexual Dimorphism. 
One of the most striking points brought out by these figures is a well-marked 
sexual dimorphism permeating all the meristic features investigated. This is well 
shown in Table 1, below, and equally for the embryos as for the adults. On 
looking down this table it will be noticed that the % adults differ from the 
TABLE 1. 
GomjKirison of the mean (M.) and its probable error (p. E. M.) betiueen the 
sexes and betiueen embryos and adidts of each sex. 
i adults 
(100) 
Antci-ior wpiiie ... 
Posterior spine 
Whole vertebrae... 
Half vertebrae 
Total segments ... 
First girdle-piercing nerve 
Post-girdle nerves 
Collector nerves 
j M. 
15-979 
\V. E. M. 
±•046 
( M. 
40-540 
(p. E. M. 
±•060 
\ M. 
44^620 
h\ E. M. 
±•077 
( M. 
40 •370 
\v. E. M. 
±•128 
j M. 
65 040 
(p. E. M. 
±•071 
) M. 
28-470 
|P. E. M. 
±•041 
! 
9-360 
(p. E. M. 
±•039 
5-230 
(P. E. M. 
±•030 
cf embrycs 
? adults 
? embryos 
(145) 
(163) 
(159) 
15-965 
16159 
16-056 
± -031 
±•040 
±•031 
40-476 
41 170 
40 686 
± -051 
± -048 
± -040 
44-510 
45 061 
44-723 
± -074 
± -057 
± •OSS 
40-034 
40-288 
39 956 
± -107 
±-101 
±•091 
64-724 
65-472 
64 906 
±•065 
± -055 
±•056 
29-057 
± -030 
7-920 
±-030 
5-304 
±-024 
The question whether any two values of a statistical constant have a significant difference in 
value is to be settled by comparing that difference with the probable error of that difference, 
which by a well-known theorem in the theory of errors is the square root of the sum of the 
squares of the probable errors of the two given values. Professor Pearson to whom I am 
indebted for this statement considers odds less than those corresponding to twice the probable 
error as not definitely significant, with odds corresponding to twice up to thrice the probable error 
we have probable significance, and with more than three times the probable error there is almost 
certain significance. Of course a difference less than twice the probable error does not prove that 
the difference is insignificant, it may merely indicate that the statistics are insufficient in number 
to adequately distinguish significant differences. Again, persistent differences of the same sign, 
when each difference is even less than the probable error, increase the odds in favour of a 
general significance. 
cf s in possessing a greater number of whole vertebrae and total segments, as well 
as in the fact that both the anterior and posterior spines are more caudally situated. 
In all these cases, except that of the half vertebrae and collector nerves, the 
difference in the means of the two sexes is more than three times the probable 
