cases secured them from error. With these principles, assisted by those of imitation and selection, the most brilliant exhibitions may be expected.

A painting by Rembrandt called the Windmill, belonging to W. Smith, Esq. M. P., and exhibited in the British Gallery in 1815, gave occasion to a conversation which induced the present communication.

A river winding round the base of a high bluff of land on which the mill stands, reflects the light of an atmosphere brilliantly white down to the horizon. The river of course is white, with scarcely any other mixture of colour.

I asked sportively of two gentlemen, philosophers of eminence, standing below it, and conversing in the room, what were the colours of seas, lakes, and great rivers; and limiting my question to sea, was answered by one, green, a mixture of yellow; by the other, the colours of Vandervelde's Sea pieces, naming and referring to them in the other room. I stated, according to the foregoing principles, that I had seen seas and waters of all the prismatic colours; seas blue, green, yellow, red; seas of quicksilver, of molten gold, of blood; waters that were black, that were white; and to that I refer you, pointing to the picture, and adding, as in the story of the cameleon, "produced the Beast, and lo! 'twas-white."

I was further led to examine the waters of the Exhibition, as represented by several masters, In Ruben's Duke of Buckingham, belonging to the Earl of Jersey, the colour of the sea is a dirty green, very much resembling the green water seen immediately from below the bridge of the Canal, not so dilute as the seas of the coast. Of Vandervelde the waters may be said to have no natural colours at all, scarcely more than the lights and shades of engravings. Cuyp's colours are such as a day of sunshine thickly overspread with masses of white and dark clouds gives to waters. This gravelly colour, which may be either external by reflection, or internal from turbid waters, Cuyp gives to his waters of Dort, with no green; and little blue.

I have stated that the colours exhibited at the surfaces of water are reflected by the air incumbent on the water, VOL. V.

H

and not by the water, which is incapable of reflecting light from or by its surface of external incidence. The general and philosophic misconceptions on this point require here to be observed upon. The existence of any power in bodies to repel light, has been formally and completely disproved by observations on the inflections of light. Light is attracted by the parts of all bodies. This is sometimes admitted and sometimes denied by observers, even to the same bodies in various circumstances. If light pass out of glass into air, and particularly when the reflection is total, this is ascribed to the glass; if partially, the reflection is assigned to the glass, the refraction to the air, by attractions of both. Let the light pass out of air into glass, both the reflection and refraction are ascribed to the glass, and in similar circumstances to water. This reasoning ascribes to the glass in this latter case, powers opposite to, and inconsistent with, those of the former case, and deprives the air of all the power in the first instance assigned to it. The air not being seen, nor so obvious as the glass, seems in this case forgotten or disregarded, and such is the state of general opinion on this subject.

At the confines of two adjacent transparent bodies, one of which at least must be fluid, spaces naturally exist different from the pores of either, in consequence of the attractions of the particles of both being different, as well as stronger for themselves than for each other; otherwise one would be dissolved by the other, as chymists know. In these spaces the bodies by their respective attractions for light, divide whatever portion thereof arrives therein into two parts, passing into and said to be refracted by the further; the other returning into, and said to be reflected by the nearer body. The nearer body can alone reflect, the further alone refract; and these are the cases of air and water in these observations. In these same spaces, and by the same attractions in other circumstances, the emission of light is produced. That these forces exist, and that by them the phenomena of the emission, reflection, and refraction of light are produced, may be proved

by arguments of the same nature and force, as those which prove the moon to be retained in her orbit by the force of gravity, by proofs establishing the existence, adequacy, and quantity, and therefore necessary agency of these forces.

ART. XII. On the original Composition of the Statues of Niobe and her Children. By Robert Cockerell, Esq.

THE statues composing the groupe of Niobe and her Children

bas long been considered amongst the first specimens of art. If all the figures were not executed by the same hand, their style and composition leave little doubt that they were the conception of one mind: they are obviously designed to form a whole, but placed without order or design as they were, and still remain, the figures appear without connexion, and to act rather in opposition to each other, than as forming a combined action or connected groupe, and they can only be regarded in their present position as single figures, without reference to their combined effect. None of the antiquarians who have noticed these statues, (including Winckelman, Fabroni, Mengs, Goethe, and Zannoni,) have attempted to solve this difficulty. Mr. R. Cockerell, whose travels we noticed in a former Number, about two years since published in Italy a plate explanatory of his ideas respecting the composition of these statues, which represents, as he maintains, the fable of Niobe and her family. This plate was accompanied by a short explanation of Mr. Cockerell's reasons in support of his opinions respecting the original composition of these statues. Although Mr. Cockerell's merits seem to have but tardily reached his own country, neither his talents nor his knowledge have remain. ed unnoticed on the continent.

These statues are supposed by Mr. Cockerell to have been originally designed for the tympanum of a pediment of a temple, the elevation and measurement of which has been

given by him in his plate, a reduced outline of which we have, by his permission, given in this Number. We shall present our readers with nearly the whole of Mr. Cockerell's observations on the subject.

The celebrated statues representing the fable of Niobe, have never been so described as to give a satisfactory idea of their relative situations, and the composition of the groupe for which they were unquestionably designed.

Montfaucon (Vol. I. p. 107,) has given a plate engraved by Perier representing these statues, ranged in a circle around the mother, as they were then placed in the villa Medicis at Rome; but this disposition, which was a mere conjecture, and entirely unsupported by the authority of the ancients, or any one single example, is entirely disproved by examination of the statues themselves, and of their different attitudes, which demonstrate that they were originally intended for only one point of view,* as will be seen from the note below, which describes the different sides of the statues.

*The statue No. 1, was designed solely for the position assigned in the groupe, for if viewed in the front, the right leg is rendered invisible by the rock which sustains it; besides which, the chest is without relief and ill executed. No. 2, on the opposite side the left leg is entirely concealed behind the rock, and the drapery suspended from the arm is but imperfectly made out. The back part of the statue No. 3, is also negligently executed and badly designed, without relief or execution. The hinder parts of No. 4, 5, 6, 7, and 8, although completely made out, are not better executed than the preceding numbers. No. 9, on the front side, the contour of the body, the hair, and the ear on the right side, are carefully finished; on the left they are merely sketched. The statue No. 10, is unfinished on the opposite side, and the right leg is concealed by a trunk of a tree. No. 11 and 12 are also left unfinished on the opposite side. No. 13: with respect to this figure, the point of view given is evidently the only possible one it could have been designed for, as the right leg is entirely wanting; and it is evident that Nos. 6, 7, and 9, were designed for a situation above the level of the eye, the different parts being more or less finished, according to the effect to be produced when seen from below.

That it was customary with the Greeks to adorn the pediments of their temples with groupes of statues is sufficiently proved by the remains of the temple of Minerva at Athens, called the Parthenon, by the discovery of those of the temple of Jupiter Panhellenius in the island of Egina, and of many other temples on which may still be traced the remains of similar ornaments, as for instance the temple of Theseus. Pausanias also has accurately described the frontispiece of the temple of Jupiter at Olympus, Diodorus that of Jupiter Olympius at Agrigentum, and many other instances might be cited.

The relative dimensions of thes statues, the gradual diminution in their height, their action, which is a general inclination towards the central point, and the simplicity and harmony of composition, resulting from their arrangement, all tend to prove that these statues were designed for the tympanum of a temple.

The passage in Pliny,* as it was written by one who may be supposed not to be conversant with technical terms, impugns what has been before laid down, but it is altogether unnecessary to cite it, as these statues may have been arranged at Rome in a manner totally different from that in which they were placed in their original situation from whence they were brought.

These statues, since their discovery in 1583, have been considered by the antiquarians as an interesting subject of discussion, both by reason of their perfect preservation, and of their extraordinary excellence as works of art; and it is singular that the authority of Ovid should have been preferred by them to that of other authors, although no circumstance of his description coincides with these statues, with the exception of the wrestlers, which are however admitted to form no

Hist. Nat. XXXVI. ch. 9. Par hæsitatio est, in templo Apollinis Sosiani Nioben cum liberis morientem Scopas an Praxiteles fuerit.