v PATTERNS .: THEIR OUTLINES AND CORES 79
The chief theoretical objection to this threefold system of classification lies in the existence of certain compound patterns, by far the most common of which are Whorls enclosed within Loops (Plates 7, 8, Fig.
12, 15, 1s, 19, and Fig. 13, 20-23). They are as much
Loops as Whorls, and properly ought to be relegated to a fourth class. I have not done so, but called them Whorls, for a practical reason which is cogent. In an imperfect impression, such as is made by merely dabbing the inked finger upon paper, the enveloping loop is often too incompletely printed to enable its existence to be surely ascertained, especially when the enclosed whorl is so large (Fig. 13, 23) that there are only one or two enveloping ridges to represent the loop. On the other hand, the whorled character of the core can hardly fail to be recognised. The practical difficulties lie almost wholly in rightly classifying a few transitional forms, diagrammatically and roughly expressed in Fig. 11, 4, 5, and Fig. 12, 8, 18, 19, with the words " see " so and so written below, and of which actual examples are given on an enlarged scale in Plates 9 and 10, Figs. 15 and 16. Here Fig. 15, a is an undoubted arch, and c an undoubted nascent loop ; but b is transitional between them, though nearer to a loop than an arch. d may be thought transitional in the same way, but it has an incipient curl which becomes marked in e, while it has grown into a decided whorl in f ; d should also be compared with j, which is in some sense a stage towards k. g is a nascent tented-arch, fully developed in i, where the pattern as a whole has a
