them, were continued, they would terminate or vanish in a point in the horizon opposite to the eye. He should next open a door about half way towards him, and at the distance of a few yards from it he will find that the nearest part will appear much too large to fit the opening, and if viewed from the other side will then seem too small. The lines of the panels, &c., will vanish as in the other case. After a few observations of this kind the pupil will begin to see things quite in a new light, and be led to reason upon the perspective appearance of the various objects around him, which will enable him the more readily to comprehend and execute such of its simple rules as it will be necessary for a landscape painter to know. All that he will require, in making his first essay, will be a sketch-book not less than eleven inches in length; and to obviate the inconvenience that will sometimes arise from the wind disturbing the leaf, he should be provided with two pieces of tin doubled, so as to admit of

being opened, to grasp the leaf and the cover; also a bit of thread with a small weight at the end of it, and a pencil that will glide freely over the paper; and as he will no doubt have frequent occasion for his India-rubber at first, it may be fastened to a string attached to the button-hole of his coat. The pupil, being thus prepared, may now attempt some little familiar scene. I will imagine, for example, that the following simple subject is before him. A thatched cottage is on his right hand, with a door in the centre, and a window on each side, and one over it. To the left there is a clump of trees, and an extensive flat distance beyond. For the foreground, a stream with a few stones in it, and some larger on its margin, with a near tree on one side. This will be sufficient for the present purpose, and which, by adhering to the following rules, it will not be very difficult to manage. The horizontal line is the first thing to be attended to, which is the line of the extreme distance against the sky that

terminates the view, and is in all cases the height of the eye. Its situation on the paper may be about one-third from the bottom, but this must depend (in drawing from nature) upon the elevation of the place from which the scene is taken. Having drawn the horizontal line, the next thing will be to sketch the foreground lightly and carefully, as the correctness of all beyond will depend greatly upon it. In order to find the situation of the cottage, he must hold up his thread so as to correspond with its nearest perpendicular, and then observe what object or part of the foreground it crosses, and over that part (previously sketched) will be the place for it, the thread being a substitute for the perpendicular line described in my first Letter. And in order to obtain the just situation of receding objects, he must imagine them to be over each other, as in a picture, otherwise his thread will be useless: and if he holds up his pencil horizontally, he will be able to judge of their relative situations on either side. Having

drawn the nearest perpendicular of the cottage, he may draw the further one at random, as to its height, and then oppose his pencil, or the edge of his book, to the lower line of the roof, as directed for the room, and where it meets the horizon make a dot on the corresponding part of the sketch, and this will be the point to which all the vanishing lines on that side of the cottage (that are parallel to each other) must be directed. The lower line of the roof will then cut the perpendicular drawn at random and reduce it to its proper height. To get the centre of the door, two diagonal lines must be slightly sketched, that is, one from the upper angle of the wall next to the roof to the opposite lower one, and then the reverse, and the point of intersection will be the centre of the wall, through which a perpendicular should be drawn to divide it (perspectively) into two equal parts, and this line will be the middle of the door; and the centre of the windows in each part may be found in the same way. I must add, that it will be particularly

necessary, in getting the inclination of the line of the roof, to keep the edge of the book or pencil perfectly parallel to the horizon, and then to slope it till it coincides with it exactly, and by no means to turn it towards the building. The distance should be drawn with a steady clean line, and the general forms of the trees in the middle distance need only be attended to; but as they advance, the stems and projecting boughs should be accurately defined with as little of touch as possible. foreground, with all upon it, should be firmly and correctly drawn, and if a portion of a young and light tree should be near enough to admit of it, the shape of a few of the leaves may be carefully marked. To get the reflection of a building, its height should be taken, and then carried downwards from its base, and this will show its situation on the water. The same rule will hold good for trees and other objects.

I am, &c.

The