ON THE PRINCIPLES OF PROPORTIONING ARCHITECTURAL ORDERS IN JAMES GIBBS'S : Rules for Drawing the Several Parts of Architecture

Abstract

James Gibbs (1682-1754) was one of most important British architects of the eighteenth century. He was also famous as a writer on architecture, and published three books, the second of which was the "Rules for Drawing the Several Parts of Architecture, in a More exact and easy manner than has been heretofore practised, by which all Fractions in dividing the principal Members and their Parts, are avoided" of 1732. The title suggested that the methodology of proportioning classical orders was the most significant feature of his book. Although the book was taken up by Rudolf Wittkower and Wolfgang Herrmann, in relation to the discussions on evolution of the dividing method in England during the eighteenth century, their arguments seem far from satisfactory, since they completely neglected Gibbs's own comments on his method. Moreover, the comparison made by Wittkower and Herrmann of Gibbs's "Rules for Drawing" with other contemporary works such as Claude Perrault's "A Treatise of the Five orders of Columns in Architecture" of 1683 (English edition in 1708 and 1722) does not seem adequate either. The present author holds that Gibbs's method of proportioning the orders and its relation to the whole system of proportion have not so far been examined to the minutest details. The present paper is to inquire into Gibbs's principles of his dividing method and consists of the following five sections : #1. Outline-of the "Rules for Drawing" #2. Method of Proportioning the Orders #3. Proportional System of Five Orders #4. Method of Dividing the Arches and Doors #5. Comparisons of Gibbs with Perrault Two thirds of the text and plates of the "Rules for Drawing" are assigned to architectural orders. Gibbs mainly deals with the proportional system of orders and the method of proportioning. As a result of his studies of the architectural books of Italian Renaissance, he intended to avoid the fractional numbers in the division of orders. In Gibbs's method, the heights and projections of components of the order are generally obtained by dividing the given length into several equal units (Figure 1 and Tables 1, 2). The 'principal parts' are at first proportioned and then divided into the 'essential parts', and finally the essential parts are sub-divided into the 'members'. Two factors of his method are important : the dividing number and the divided ratio. The dividing number is a number into which a part is divided in height and projection. As a result of the division, a component can be measured by its ratio to another component. This is the divided ratio. The ratio is a combination of integral numbers, the sum of which is equal to dividing number (Table 1, A). The idea of 'the method of dividing the orders Mechanically into equal parts' seems to be well materialized here. But there are cases where the sum of integral number which the ratio consists of is not equal to the dividing number (Table 1, B and C), because the measurements of components are taken by adding the unit of the first division to that of another division. In such cases, the measurement is not divided mechanically. Thus, Gibbs's method of proportioning the orders contains two contradicting approarches. It is possible to transform the latter method into the former, but the dividing number becomes larger. Therefore, Gibbs must have introduced the second method in order to keep the dividing numbers small and the divided ratios various. The dividing numbers in Gibbs's system are smaller than ten and Gibbs always adopted two or three for the vertical division (Table 3). As for the horizontal division they are below ten and two is usually taken. There is no apparent rule in their usage, except that the first stage of division begins with five for all the orders and the maximum numger is used in

(View PDF for the rest of the abstract.)