The area of the most acute vision in the horse is probably the "tapetum," and that portion of the tapetum which lies in the optical axis of the eye is the part where all the observations contained in this paper have been made.

The horse to be examined is brought into an ordinary stable, and any windows likely to be reflected in the cornea are covered up. The observer stands facing the source of daylight, the eye to be examined being in shadow; standing opposite to the centre of the pupil (and looking at it in such a direction that a line passing through the eye would come out well below the base of the opposite ear), the light is reflected on to the retina at a minimum distance of 4 feet. A brilliant yellow reflex is obtained from the tapetum, by means of which we can readily fix our position relative to the pupil.

No means of restraint are employed, and though for a minute or two the horse works the eye under examination to and fro, thereby alternately altering the reflex, yet it soon comes to a standstill as he gets used to the reflection from the mirror.

The observer, by practice, automatically regulates his position, so as always to be opposite the centre of the longitudinal pupil, and to ensure that during any change he may make in his position he does not come nearer to the eye than the prescribed distance, a 4-foot gauge is lying on the floor at his feet, in the direction of the horse's head.

The reflex is taken in the horizontal and vertical meridians of the eye; it is very seldom that any difficulty is experienced in deter mining the direction in which the shadow is travelling; the real difficulty is in determining between two opposite shadows given consecutively in the same eye and in the same meridian. I think this latter phenomenon is caused by the horse using its ciliary muscle; with patience the true reflex of the resting eye will be obtained.

Having determined the direction of the shadows in both meridians, a trial lens is now placed in a frame and held in front of the eye under examination, The glass is brought quite close to the cornea, the centre of the lens corresponding to the centre of the cornea; the observer, standing at a minimum distance of 4 feet, throws the light into the eye and determines the reflex through the lens.

As we approach the lens which corrects the ametropia, some diffi culty may be experienced in determining the direction taken by the shadow; sometimes it neither passes "with" nor "against" the mirror, but dies away from the circumference, producing an appear ance resembling a dissolving view. In human practice this charac teristic reflex, I believe, is considered to be due to astigmatism.

No part of the above examination is made under atropine; as previously explained, the pupil as a rule dilates sufficiently for a complete examination of the fundus to be made, and, so far as

paralysis of accommodation is concerned, little, if any, is produced even by repeated instillations of this drug.

My results have been calculated according to the rule given by Morton: For hypermetropia deduct +1 D from the weakest convex lens which causes the shadow to move with the mirror; for myopia deduct +1 D from the strongest concave lens with which the shadow continues to move with the mirror.

The following table (p. 420) exhibits the results of the enquiry. The table shows that of the 100 eyes examined

The varieties of astigmatism which existed were as follow:

Compound myopic astigmatism

43
8

In the eyes affected with myopia and astigmatism, the horizontal meridian was nearly always the meridian of least curvature. Berlin and Lang and Barrett had also observed the same fact. In my table of the 51 eyes there were only nine exceptions to this rule.

The amount of error which existed in each refraction was as follows::

Only once in 39 eyes was the amount of 3 D reached, whilst the most common amount of error was 0.50 D.

VOL. LV.

2 G

The most common amount of error was 0·25 D and 0·75 D.

The amount of error shown in these different refractions is small, so much so that many eyes might be regarded as emmetropic. In only about 33 per cent. of the myopic eyes did the amount of error exceed 1 D.

The number of astigmatic horses is remarkable, viz., 74 per cent., but we cannot agree with Hirschberg that the amount found is "intense." Of 59 astigmatic eyes only in nine did the error exceed

1 D.

The following horses in my table gave a history of "shying," viz., Nos. 9, 13, 18, 19, 25, 26, 27, 28, 32, 42, 48, and 54; from this list I have excluded three doubtful cases, but it must not be supposed that all the above horses "shied" so badly as to constitute a vice; nor

must I be understood to say that all shying is due to defective eyesight; further, it is possible for defective vision to be present without shying occurring, notably in Case 23 and several others.

Berlin considered the majority of horses to be hypermetropic; I have only met with one case of very low hypermetropia out of 100

1. The chief visual defect in horses is myopia with or without astigmatism.

2. The amount of error is not great.

VI. "Correction of an Error of Observation in Part XIX of the Author's Memoirs on the Organisation of the Fossil Plants of the Coal Measures." By W. C. WILLIAMSON, LL.D., F.R.S. Received March 28, 1894.

During the past two or three years we have acquired a much more accurate knowledge than we previously possessed of the structure of the leaves of the Lepidodendroid plants of the Coal Measures; and my last memoir, published in vol. 184 (1893) of the Philosophical Transactions, contains some new and not unimportant details on this subject. These facts were obtained from some finely preserved specimens of Lepidodendron Harcourti and of a Lepidophloios recently added to my cabinet.

But the value of these observations was seriously vitiated by an apparently small but very fundamental error into which I had fallen, an error which cannot be left uncorrected without misleading some of the paleontologists who honour me by consulting my writings on these subjects.

The students of fossil botany have long distinguished the leaves of the Lepidodendron from those of the Genus Lepidophloios by the shape of the leaf-scar left on the pulvinus or leaf-cushion on the fall of the deciduous leaf. These scars are always more or less quadrilateral in form, two of their angles following in opposite directions the long axis of the parent stem or branch, the other two pointing transversely across that axis. In Lepidodendroid leaves there is little difference in the lengths of these pairs of decussating angles, but in Lepidophloios the two transverse ones are much more prolonged than the vertical ones are, making the transverse diameter of the leaf-scar greatly exceed that in the opposite direction.

Whilst attached to the stem or branch of a Lepidodendron these leaves always point upwards towards its apex; and when I wrote the memoir referred to above, I had no reason for supposing that this was not also the case with the foliage of Lepidophloios, and my two