Here (Fig. 9) is a cone poised on its apex, its axis vertical: as a simple illustration propoitionately it is contained in a square, call it a foot. A A is the axis, the point P proportionately three inches from the top or one-fourth down the axis. This should be your first measurement. The next thing is to fix your diameter at right-angles to the axis and equidistant on each side. (A mnemonic here is the letter T, which will serve to remind you that the diameter must in all cases be made at right-angles to the axis.)

Fourthly, the ellipse which the circle in perspective becomes should be considered, and practice is very necessary to attain freedom in its drawing. To obtain balance, draw opposite curves through the points, not to them, the pencil, in plan, as nearly as possible perpendicular to the line (see Fig. 10). 1 - 2, 3 - 4, 5 - 6, 7 - 8. Note that from pole to pole is a changing, living curve: immediately on reaching a pole the line changes alternately from "flatish" to roundish, from roundish to flatish; a true geometric ellipse, which might, by folding on the lines 1 - 2, 3 - 4, bring all four parts into coincidence. This, theoretically, may be questioned, but in practice will be found very accurate indeed. As one diameter (the minor axis) of the ellipse is included on the axis of the solid, it is found more convenient to refer to the major axes of ellipses as diameters throughout the article.

After the ellipse is drawn the contour may be added. It consists of two lines drawn from the apex tangentially to the curve of the ellipse. A common error is to draw to the extremities of the diameter. You have now the cone complete, and on this simple model may be built a system for the drawing of all conical and cylindrical objects, which is here denoted as the Axial System (Fig. 11) in this order:

1. Axes.

2. Proportions.

3. Diameters.

4. Ellipses.

5. Contours.

Condensed for remembrance to the letters A P D E C (which looks like Volapuk, but isn't) and you have your system at your fingers' ends (Fig. 12).

A glass cylinder- (or a clear plain tumbler inverted), which can be seen through, would be the best to demonstrate the method here considered. In Fig. 13 two proportions are marked off on the axis, the lower one deeper than the upper.

A circle facing us is seen as a circle; on edge as a straight line; in all other positions as an ellipse. In Fig. 13, the upper ellipse has a shorter proportion on the axis, because it more nearly approaches to a position when the circle would be seen on edge - deeper below, because it more nearly approaches the circular view. Similarly, if viewed when vertical, circles grow wider left or right, and if we had 100 they would get wider as each was removed farther from the central one viewed on edge (see Fig. 14). With this additional information it should not be difficult to complete the cylinder. To gain proof of accuracy here, the student may cut his drawing in half, and, holding one half of it between the eye and the object, the edge of the paper coinciding with the centre of the cylinder, and seeing that the half drawing pairs exactly with the half cylinder at proper locus, a valuable point in visual training is attained.

It is also one of the most convincing experiments, and may be applied to any position (see Figs. 15 and 16). In dealing with positions other than vertical, the student will do well to ask himself the question in relation to the axis, " What o'clock is it?" This will give him a very approximate idea of the angle of the axes in all possible directions if faithfully carried out. Holding the pencil before the eye, imagine it as the hand of a clock passing round to the moment when the axis occurs; fix the axis and follow A P D E C (Fig. II). A thousand familiar tonus may be drawn on this plan with very slight additions and variations of contour. In Fig. 17 three different contours are indicated on a cylindrical base to show effects of change of contour. One hint more may be added, namely, that the sections of contour are tangents to the ellipses. Edward Renard, A.R.C.A. (Lond.).

(To be concluded.)

Although fruits are often more brilliantly coloured than flowers, and their shining rinds are more difficult to reproduce than the mat texture of petals, they are nevertheless easier for the beginner because of their solidity, which makes the modelling much simpler, and their roundness, which permits of every touch being run into the preceding one. There is no such trouble as is given by the varied and capricious curves of flower petals, their difficult foreshortenings, multiple reflections, lights, shadows, forms and tones more or less obscurely evident through their substance. Fruits are real solids, and most of them opaque. It is in their favour, too, that they last much longer than do flowers.

Object Drawing for Craftsmen. By Edward Renard, A.R.C.A. (Lond.).