I have already shown that such figures as the scroll and the guilloche are not _necessarily_ developed by processes of selection and combination of simple elements, as many have thought, since they may have come into art at a very early stage almost full-fledged; but there is nothing in these facts to throw light upon the processes by which ornament followed particular lines of development throughout endless elaboration. In treating of this point, Prof. C.F. Hartt[2] maintained that the development of ornamental designs took particular and uniform directions owing to the structure of the eye, certain forms being chosen and perpetuated because of the pleasure afforded by movements of the eye in following them. In connection with this hypothesis, for it is nothing more, Mr. Hartt advanced the additional idea, that in unison with the general course of nature decorative forms began with simple elements and developed by systematic methods to complex forms. Take for example the series of designs shown in Fig. 488. The meander _a_ made up of simple parts would, according to Mr. Hartt, by further elaboration under the supervision of the muscles of the eye, develop into _b_. This, in time, into _c_, and so on until the elegant anthemium was achieved. The series shown in Fig. 489 would develop in a similar way, or otherwise would be produced by modification in free-hand copying of the rectilinear series. The processes here suggested, although to all appearances reasonable enough, should not be passed over without careful scrutiny. [Illustration: FIG. 488.--Theoretical development of fret-work.] [Footnote 2: Hartt: Popular Science Monthly, Vol. VI, p. 266.] Taking the first series, we observe that the ornaments are projected in straight continuous lines or zones, which are filled in with more or less complex parts, rectilinear and geometrically accurate. Still higher forms are marvelously intricate and graceful, yet not less geometric and symmetrical. [Illustration: FIG. 489.--Theoretical development of scroll work.]