beautiful engravings by Lehmann after the drawings by the author.

Chatin: Anatomie Comparée des Végétaux. Good steel engravings illustrating the structure of various plants.

Curtis: Flora Londinensis, London, 1777. The illustrations are hand-coloured copper engravings by Sowerby and others, many of which, particularly the earlier ones, are of outstanding excellence. The engraving is often nothing more than the mere outline of the plant, whilst in cases where the structures are more massive, a certain amount of shading is used. The colouring is very good indeed, and it is obvious that much care was taken not only in the actual painting but also in the choice of pigments which, as far as can be judged, are as fresh now as when first used.

Curtis's Botanical Magazine and Edwards's Botanical Register contain some excellent examples of hand-coloured copper engravings.

Levaillant: Histoire Naturelle des Oiseaux d'Afrique. Paris, 1805-8. This work contains beautifully coloured engravings by Feesard. The original drawings were by Reinold.

Lyonet: Traite Anatomique de la Chenille. La Haye, 1762. The plates are amongst the best illustrative of zoological science.

Martius: Flora Brasiliensis. The earlier volumes, e.g., Vol. 13, Part I, contain excellent engravings.

Passæus: Hortus Floridus. Arnheim, 1614-17.

Sowerby and Smith: English Botany. London, 1790-1866. The illustrations are hand-coloured copper engravings.

Thuret et Bornet: Etudes Phycologiques. Paris, 1878. This work contains the finest plates ever published in a botanical work. Riocreux drew from the preparations, and his drawings were engraved on steel by Picart, Thomas and others.

ETCHING. Etching is a term very loosely used; strictly speaking it consists in corroding a metal plate or a flat stone with acid, or other substance possessed of a kindred action, so that depressions are formed. A pen and ink drawing, although usually so termed, is not an etching. Briefly the method is this: a well polished copper, steel or zinc plate is covered with a substance, known as the etching ground, consisting commonly of a mixture of asphaltum, white wax and pitch, which resists the action of the acid. The ground may be laid in more than one way; the simplest, perhaps, is to dissolve the etching ground in some solvent such as chloroform, which readily volatilises, and to pour the solution on to the plate, which is tilted this way and that until the liquid is evenly distributed; the excess is poured off and what remains is allowed to dry, the plate being kept level during the process.

The plate is then warmed until the ground is softened, when it is held over a smoking candle and is rapidly moved here and there so that if properly done the fine soot is evenly incorporated in the ground. When the plate is cold, the drawing may be made by cutting through the etching ground, so as to expose the underlying copper, with needles of various sizes. The work is then etched by means of dilute nitric acid.

When this is satisfactorily accomplished, the ground is cleaned off, the plate well inked with copperplate ink, and the surface ink removed by coarse muslin. The plate is then gone over with fine muslin, but the ink must not be removed from the depressions; finally the damped paper is placed in position and impressions obtained by the use of the copper-plate press.

Etching, although suitable, especially when natural-printed,* for the illustration of many scientific subjects, is but seldom employed at the present time for this purpose; the preliminary announcement of Warburg's Die Pflanzenwelt, however, states that some of the illustrations are etchings.

*A plate is said to be natural-printed when all the ink is removed except from the depressions; in artificial printing some ink is allowed to remain on the flat parts. Artists frequently, after removing the superfluous ink, lightly dab the plate in order to make the pigment spread slightly beyond the actual limits of the depressions; this is known as retroussage.

SOFT-GROUND ETCHING. This is a somewhat rare method of reproduction nowadays; it may, however, be described briefly, for it would appear to be suitable for scientific purposes, since it should not prove a matter of great difficulty for an author who is a sufficiently skilled draughtsman to make his own plates. The polished copper plate is laid with ordinary ground to which is added lard in a quantity according to the warmth of the weather.

Over the plate is then placed a sheet, larger than the plate by an inch or two, of damp, thin, grained paper, the edges of which are folded over and pasted to the back of the plate. When the paper is dry it will be well stretched and in close contact with the plate. With the hand resting on a bridge, in order to avoid inadvertent touching of the plate, the drawing is made on the paper with a pencil of a hardness suited to the softness or otherwise of the etching ground. When the drawing is finished the paper is carefully removed; wherever the pencil has been used, the etching ground will adhere to the paper, so that in such places the metal will be exposed. The plate is then etched and printed as in the normal process.

No reproductions of drawings of scientific subjects apparently have been reproduced by this method. Examples can conveniently be examined in The Seven Lamps of Architecture by Ruskin.

MEZZOTINT. The characteristic feature of mezzotint is that the subject is translated into tones rather than lines as in the preceding intaglio methods.

The surface of a smooth metal plate—usually copper—is raised into innumerable and minute projections by going over it in all directions with a curved steel tool, known as a rocker, the edge of which is finely toothed. An impression taken of the plate in this condition will give a deep rich tone. The high lights are obtained by scraping and burnishing away the elevations so that there are no pits left to hold the ink, and, similarly, intermediate tones are produced by partly removing the pile so that the pits are made of varying degrees of shallowness and consequently will print in tones according to their depth.

Impressions are taken in the same way as in the case of etchings.

Mezzotint apparently has never been used for the reproduction of scientific subjects. Indeed, in a sense, this process is much too artistic for the purpose. At their best, illustrations reproduced by this method have mystery and depth and give the imagination much employ; in a word, they are subjective rather than objective, qualities unsuited for our purpose.

PHOTOGRAVURE. Photogravure may next be considered, for although it is a photo-mechanical process, it corresponds to mezzotint. Excellent results may be obtained by its use provided the drawings—usually executed in monochrome such as sepia—be really good, otherwise they are hardly worth reproducing by this relatively expensive method.*

*This account refers only to plates made and printed entirely by hand, not to photogravure for printing on a rotary machine.

Photogravure is particularly suitable for the reproduction of drawings showing a large amount of detail and made up of a variety of tones rather than lines or stipple.

The photographic part of the process is essentially the same as making a carbon print from a photographic negative. This consists in exposing under the negative the carbon tissue, which is a mixture of gelatine, in which is dissolved bichromate of potash, and a