246 VARIATION, DISTRIBUTION, AND EVOLUTION OF THE GENUS PARTULA. 
Analyzing the material of the second line, on the assumption that banding 
is recessive, the proportions of the genetic classes are as 8:12:13, which give a 
departure of only 1 in 56, or 1.8 per cent in the corroborative test. Supposing the 
contrary, the proportions are the less satisfactory figures, 9:13:34, with an indi- 
cated difference of 6.06 per cent in the further analysis. 
RESUME. 
The general results, combined in table 191, lend considerable support to the 
contention that the unbanded pattern is dominant to the banded form of colora- 
tion in a Mendelian order of inheritance. In view of the many difficulties involved, 
it is not to be expected that the results would be more consistent than they prove 
to be in this case. 
Tas Le 191.—Partula otaheitana sinistrorsa. Summary of the results of a Mendelian 
analysis of the color-patterns in heredity. 
Assuming plain as dominant. Assuming banded as dominant. 
Proportions of | Difference in Proportions of Difference in 
DD :DR:RR.|independent test.; DD : DR: RR. |independent test. 
Tenaire: pb. ct. 5 Ae 
GCombinedmererece: 31: 85: 41 Bt Sil g Sg Sew i777 
Sinistralfererereeeecee ill g 39g Oo) oe) N83 Ae 8 ils 2.8 
DMextralivac. soeiecce 20: 47: 20 1.05 Impossibility. |  ..... 
Maara: 
Combinedireeeeaee 8: 20: 34 26.6 22 2338 2 30.6 
Sinistral aan G8 ile 28 34.4 Lg leg 1 27.5 
Dextral here Ae Og ili Sez I'm possibility-"| ieee 
Faarahi, sinistral..... 149 Bg Ai Uo iS) 9 S39 OY 29.5 
Moaroa, sinistral..... 8 179 is 1.8 9 1S 8 S¥h 6.06 
SUMMARY AND CONCLUSION. 
The series herein described as Partula otaheitana sinistrorsa grades so directly 
into P. o. sinistralis, and this in turn is so similar to P. 0. crassa, that a summary 
discussion might well be deferred until all of the colonies as far as Taapuna had been 
described. In that case, however, there would be 34 associations to be considered. 
It is preferable, therefore, to consider the three divisions separately, always empha- 
sizing their closer relationship, which in the case of the sinistrorsa-sinistralis transi- 
tion almost amounts to continuity. 
I. The colonies of P. 0. sintstrorsa vary from valley to valley (1) in their relative 
numbers in the whole population; (2) 1n the proportionate numbers of sinistral and 
dextral individuals, when the latter occur; (3) in the relative numbers assigned to the 
several color-classes, and (4) 1n details of color-variation within one or another color- 
class. It is unnecessary to repeat at length the statements of earlier general dis- 
cussions as to the significance of such variations. While the Taiarapu areas may 
differ ecologically to some slight degree, the valleys of southern Tahiti nui are 
similar in every observable respect; in no case are there discoverable environmental 
factors that might account for the special features of a sznistrorsa colony in a single 
