58 
A Study in Variation. 
within that group. The scatter diagrams show the correlation 
or otherwise of two variants, a wide and indiscriminate scatter 
indicating low correlation, while the location of the scatter 
in two diagonally opposite quadrants indicates a high degree 
of correlation. As a further measure of correlation the 
arithmetic mean deviation has been calculated, and also 
expressed as a percentage of the mean measurement of the 
variant, in the same way as indicated by Vernon (i). The 
straight plotting of height and breadth as ordinates of a scatter 
diagram enabled the limiting curves — a maximum and a 
minimum — to be calculated, thus giving at any point the 
limits between which any specimen might be expected to 
vary. 
Consideration of the Frequency Polygons. — In all 
the diagrams the solid lines represent the polygons in which 
the six hundred specimens are considered. In all cases the 
boundaries of the polygons are full of minor irregularities 
due to the lack of numbers ; however, in three of the four 
diagrams (Fig. 3, B.C.D.) two major peaks are observed, the 
exception being X 100 in which case the polygon gives 
impression of homogeneity, being a normal binomial curve. 
Judged on this diagram alone there is nothing to suggest any 
abnormality in the group, this phenomenon being due to the 
fact that the variants change in the same way and increase 
at similar rates. 
The three polygons with irregularity each show one large 
peak on the right and a smaller one on the left. When this 
fact was discovered the specimens falling within the two areas 
enclosed by the peaks were separated and examined further. 
Three characters were found to differ, first the anterior margin, 
second the tumidity, and third the sharpness of the ribs. 
There were smaller and younger specimens of the two species 
somewhat intermingled, and in this preliminary sorting it 
was necessary to fix on one difference between the two. Since 
the internal characters showed no obvious difference, the 
character of the anterior margin was taken as the feature of 
differentation, with the result that 450 specimens fell into 
the species Pugnax pugnus Martin, and 150 into the group 
which is here named Pugnax cf. sulcatus J. de Sow. The 
differences are seen in the drawings Fig. 2 where the anterior 
margin, the tumidity, and the sharpness of the ribs constitute 
marked differences in the species. 
Variation in Pugnax pugnus Martin. — Martin (2) gives 
the type species and figures it in Plate 22 of his work as 
Conchy liolithes anomites subrotundus. The drawing of the 
anterior view is not copied by later writers. Davidson (3) 
omits the layered nature which Martin shews and which is 
very characteristic. The species is described by Davidson 
The Naturalist 
