342 PROCEEDINGS OF THE AMERICAN ACADEMY 



result from the pressure of the architrave upon the pliant stems.* This 

 explanation would he complete if it also showed why the bulge is at 

 the bottom of the capital rather than at its middle height. The latter 

 form seems to have been seen to be ungraceful, and abandoned for one 

 which should in some measure repeat the figure of the shaft. It is 

 also not impossible that the analogy of the true bud capital may have 

 influenced this form. 



If the second sub-group of papyrus columns be not wrongly inter- 

 preted, the preceding theory should apply to it also. This it does, I 

 think, with special success; for although this capital has been under- 

 stood to represent a single bud, its form is really not at all like a bud, 

 — the expansion is too low and too abrupt. The point of greatest 

 breadth seems to have been determined after the analogy of B. l.f 



The only decorations vouchsafed either of these capitals are sheaths 

 similar to those found at the foot of the shaft. These, of course, are 

 confined to the papyrus columns, where they are almost invariable. 

 They regularly enfold the projecting edges of the stems, and alternate 

 with the astragal pieces, under which they are partially concealed. 

 (See cuts.) Their use can hardly be traced to any natural type. It 

 was probably suggested by the sheaths on the shaft. 



The relation between the top of the capital and the abacus is differ- 

 ent with the two kinds of columns. The thin plate which surmounts 

 the lotus columns extends considerably beyond the top of the capital ; 

 but the heavy block of the papyrus columns, with one exception, t con- 

 forms exactly to the dimensions of the top of the capital. 



The proportions of the capitals I was able to compare are altogether too di- 

 verse to admit of any general theory concerning them. 



The height of the capital, measured by the shaft-height, is about .22 in A., 

 and from .28 to .38 in B. The same, measured by the column-height, is .17 in 

 A., and from .20 to .25 in B. Again, measured by its own greatest diameter, 

 it is 1.20 in A., and from .92 to 1.23 in B. 



The greatest diameter of the capital in A. almost exactly equals the greatest 

 shaft-diameter, but in B. it varies from .90 to 1.13 times that diameter. 



In about half the examples, the greatest and least diameters of the capital are 

 nearly proportional with the corresponding diameters of the shaft. The remain- 

 ing examples depart widely from this proportion. 



* Wathen, Arts, etc., pp. 98, 109. 



t Taking good examples of the three varieties, I find that in A. this point is 

 about one fourth of the capital-height from its base; in B. 1, one seventh; and 

 in B. 2, one twelfth. 



f. From the Fayoum. Lepsius, i. 47. 



