Ethel M. Elderton 
65 
Now if we treat our four fundamental types as distinct we have the f iUowing 
results of crossings : 
Arch X Arch = 9 Arches + 6 Loops + 1 Composite + 0 Whoils, 
Arch X Loop = 56 Arches + 123 Loops + 17 Composites + 13 Whorls, 
Arch X Composite • = 5Ai-ches+ 17 Loops + 4 Composites + 0 Whorls, 
Arch X Whorl = 7 Arches + 40 Loops + 4 Composites + 1 2 Whorls, 
Loop X Loop = 39 Arches + 288 Loops + 28 Composites + 59 Whorls, 
Loop X Composite =16 Arches + 77 Loops + 18 Composites + 12 Whorls, 
Loop X Whorl = 26 Arches + 180 Loops + 37 Composites + 100 Whorls, 
Composite x Composite = 0 Arches + 6 Loops + 2 Composites + 4 Whorls, 
Composite X Whorl = 0 Arches + 12 Loops + 12 Comj^osites + 22 Whorls, 
Whorl X Whorl = 0 Arches + 16 Loops + 6 Composites + 24 Whorls. 
These are the equations which the Mendelian has to intei'pret. Certain things, 
are at once obvious : 
(i) That both Arch x Arch and Arch x Composite appear to give no whorls. 
(ii) That both Whorl x Whorl and Whorl x Composite ;ippear to give no 
arches. 
(iii) That Arch x Loop, Whorl x Loop, Composite x Loop and therefore not 
unnaturally Loop x Loop can reproduce all four types. 
(iv) That the Composite is so relatively rare that it is difficult to lay stress on 
the numbers obtained fir it. But Composite x Composite suggests on meagre evi- 
dence that this mating cannot produce arches. 
(v) That Arch x Whorl can reproduce all types. 
These results suggest that the loop is of a much more heterozygous character 
than either arch or whorl. If we throw composites in with whorls, the Composite x 
Whorl by no means gives the same result as Whorl x Whorl ; and if we throw 
composite in with loop we do not find the etpiations for Composite x Loop, Arch or 
Whorl in satisfactory agreement with those for Loop x Loop, Arch oy Whorl. The 
evidence, meagre as it is, seems to indicate a real genetic differentiation between 
Composite and Loop. 
Still if we venture to put Composite with Loop, we obtain a much .simplified 
series of equations, which admit of being expressed as follows in round numbers : 
Ratios 
(i) Arch X Whorl = 6 Arches + 42 Loops + 12 Whorls, (1:7:2) 
(ii) Arch x Arch = 30 Arches + 30 Loops + 0 Whorls, (1:1:0) 
(iii) Arch x Loop = 16 Arches + 40 Loops + 4 Whorls, (4 : 10 : 1) 
(iv) Loop x Loop = 6 Arches + 45 Loops + 9 Whorls, (2 : 15 : 3) 
(v) Loop X Whorl = 4 Arches + 38 Loops + 18 Whorls, (1 : 10 : 4) 
(vi) Wh(jrl X Whorl = 0 Arches + 30 Loops + 30 Whorls. (0:1:1) 
Biometrika xiii 5 
