F. G. Percwal—Punctation of Terebratulid Shells. 538 
the sides, and one travels over the stages from youth to age more 
quickly in an oblique direction than along the median line. The 
variation is negligible along any particular zone of growth parallel 
to the growth-lines. 
It follows from the foregoing that one must always choose some 
fixed part of the test for the counting. In the earlier work the 
valves were divided into squares of one centimetre, commencing 
at the umbo, the lines being painted in water-colour. Readings 
were taken in the second square, i.e. at between one and two centi- 
metres from the umbo, on the dorsal valve when possible, otherwise 
on the ventral. The figures for the ventral valve are approximately 
the same as for the dorsal. 
If the density'is to be of any value the range for a species must 
not be very great. But it appeared in one or two cases that there 
was a rather large variation. Moreover, as previously mentioned, 
occasionally a sudden change occurred. For example, a specimen of 
Microthyris sublagenalis, Davidson, showed in the first square centi- 
metre a band where the density rose suddenly from 168 to 272 (see 
Pl. III, Fig. 2). A strong growth-line separated the two readings, 
but farther out, in the third square from the umbo, the density had 
fallen to 152. Obviously, if irregular variation like this is at all 
common there will be a big range for each species, even if we confine 
our readings to a spot of very small area at a fixed distance from the 
umbo in each example. 
In order to test this large numbers of two species were collected. 
The first was 7. beplicata “(Brocchi), using the term in Davidson’s 
broad sense, of which 166 examples were obtained from the 
Manchester Museum collection. All were from one horizon and 
loeality—Lower Chalk, Sewell. There was a fairly wide variation in 
form, but there was a complete gradation among them. Over 
300 specimens of 7. punctata, Sow., were also collected from a couple 
of blocks in a Middle Lias quarry at Stathern, Leicestershire. In 
these specimens the puncte were usually beautifully preserved and 
yet difficult to count. Staining did not improve them at all. Most 
of them had transparent shells with a black matrix inside and filling 
the puncte. The matrix showed through the test, and the puncte 
were almost invisible against the dark background. On heating the 
fossils on wire gauze over a Bunsen flame the tests cleaved into 
myriads of tiny rhombs, giving an opaque white appearance against 
which the black puncte showed very distinctly. After heating, the 
surface was covered with Canada balsam to prevent flaking, but in 
some cases it was an advantage to let the thin outer layer flake off, as 
the round cross-section of each punctation in its inner half made 
a very distinct spot, easy to count. 
At first photomicrographs of known magnification were taken, but 
this was too slow a process. The later work was done by using 
a long extension camera, giving» an image magnified fourteen 
diameters. A square of 14 millimetres was cut out of a sheet of 
paper, which was then moved into various positions on the screen. 
With the aid of a lens several readings were thus taken for each 
individual, by counting the numbers in the square. These readings 
