reversing the combination in the V-shaped mount, refocussing the camera and varying the direction of the daylight illumination, a very successful negative was secured, a reproduction of a positive from which is seen in fig. 2 (Plate 2). Around the centre of the lens will be observed a great number of small white spots which represent the markings in question. These are quite different in form and character from any of those which have been previously described.

In the case of two 4-in. lenses a recent examination has shown that after an interval of 35 months they are just beginning to show signs of development of markings.

In this initial stage one has observed, with the help of a watchmaker's glass only, that the markings take the form of very small single streaks, more or less homogeneously scattered over the whole lens, but a little more sparse nearer the edges than about the centre.

The large 12-in. objective has developed those markings at a much greater rate. After an interval of only 23 months very fine markings, of simple structure, but fairly closely packed together, extend over the whole surface of the lens, and, as before, their structure becomes more complex as the edge of the lens is approached.

So far, no mention has been made as to the particular surface of the combination on which the markings are to be found. As a matter of fact, although it is quite simple to take to pieces any of the lenses which have been examined, it is not easy to replace them properly centred in their cells. For this reason, therefore, only one of the above-mentioned lenses has been thoroughly examined by separating the component lenses forming the combination. The particular objective thus treated was that of 3.5 inches aperture, and it exhibited both kinds of markings to a very great degree.

When markings were first noticed to develop on one of the inside surfaces of the combination it was thought that they might be due to the possible lack of permanency of the boro-silicate flint glass, since there seemed more doubt connected with the permanency of this material* than with those used for the other lenses. It was therefore found, with some surprise, that when the 3'5-in. objective was taken to pieces, both surfaces of the negative lens were perfectly clean and unmarked, and therefore not deteriorated in any respect. This result was highly satisfactory from the point of view of the lasting power of the boro-silicate flint glass, and showed that, as mounted between the two other lenses, 40 months' use had not affected it. It was the inner surface of the back lens, of light silicate crown glass, which was affected to the greatest extent, and on which the markings illustrated in the above figures were scattered. This surface, it will be remembered, is separated from the back surface of the negative lens by a small air-space. It may be mentioned that no balsam or like Especially where sulphuretted hydrogen or sulphuric acid abounds, as in a town atmosphere.

*

material is employed in this combination to cement the lenses together, but the surfaces are kept separate by thin washers.

In order to investigate more thoroughly the nature of the two different kinds of markings, namely, those which appeared on the inner surfaces of the back and front lenses of the 3.5-in. objective, the combination was taken to pieces and the markings microscopically examined.

This inquiry was facilitated by the kindness of Professor J. B. Farmer, D.Sc., F.R.S., of the Royal College of Science, who not only allowed me to use one of his microscopes in his own room, but also examined the markings with me. It was at once obvious that we were not here dealing with any form of vegetable growth, as seemed at first probable by the general appearance of the markings, but that both kinds were of a crystalline nature.

In the case of the back lens the crystallisation seemed to have taken place on the surface of the lens, and had the appearance of a deposit on the surface. Each fungus-like figure seemed to have originated from a small area, not a single point, and developed radially in numerous directions, throwing out ramifications as the growth from this area progressed. Under the microscope it was found that the markings were really crystals, forming at first about the small area, and then gradually growing outwards, crystal by crystal, until they formed chains of crystals, giving the fungus-like growth described above. There seemed little doubt that if this process were allowed to continue, the individual groups of crystals would all eventually unite, and seriously interfere with the transparency of the combination. In the lens under examination several of these groups have already become interlaced and instances of this can be seen in the centre portion of the lens illustrated in fig. 1 (Plate 2).

In the case of the markings on the back surface of the front lens, the form of crystallisation is quite different. Here, under the microscope, the markings (seen in fig. 2, Plate 2, as white dots scattered about the centre of the lens) are composed each of a little irregular group of crystals, formed one above the other, and projecting from the surface of the lens, not embedded in it. There is little appearance of lateral growth to the extent described in the case of the markings on the back lens. Scattered over the surface of this lens are also small single crystals, which suggest the first stage in the building up of these heaps. In this case also it is probable that if this process of crystallisation were allowed to continue, more single crystals would be formed, the single ones already existing would each develop into a small pile of crystals, and the lens would gradually become less transparent.

By the kindness of Mr. W. H. Merrett, A.R.S.M., the Instructor in Assaying in the Royal College of Science, I was permitted the use of the photomicroscope set up in the metallurgical laboratory of that institution. With the help of this instrument and with his assistance I was able to take large scale photographs of the

'crystalline formations on each of the two lenses, and some of these are reproduced on Plate 3.

Figs. 1, 2, 3 show three of the individual small groups of crystals formed on the back surface of the front lens, to which reference has just been made. The magnification employed for these photographs was 700. It will be seen that the formation is quite irregular, and, as a rule, there does not seem to be a point of origin, as if their existence was due in the first instance to the presence of a speck of dust. On the surface of this lens there appear also single crystals which are totally different from those just described. They do not show the lateral growth, as is here illustrated in figs. 1, 2, 3, but they are very much more raised from the surface on which they are formed.

The markings on the inner surface of the back lens of the combination are illustrated in figs. 4, 5, and 6 of Plate 3. Fig. 4 gives a general idea of the middle portion of one of the groups under a magnification of 45 diameters. It will be noticed that there does not seem to be a distinct single point of origin, but that the group seems to spring from a small area. An interesting point to which attention may be called is the fact that none of the lines of crystals cross each other: this can be observed in several instances in fig. 4 (Plate 3) where these lines approach each other. Further, the extremity of each line, where the crystallisation can proceed unhindered, is nearly always capped by a complete crystal.

Fig. 5 (Plate 3) is a more highly magnified (150 diameters) portion of fig. 4, and for identification purposes similar portions in figs. 4 and 5 have been lettered A, B, C. In this figure the building up of the lines by successive deposits of crystals can be more clearly seen.

Fig. 6 (Plate 3) is a portion of another group altogether. The magnification here is 700 diameters. It shows perhaps more clearly not only the individual crystals which help to build up the lines which compose the groupings previously mentioned, but the complete form of the last crystal deposited at their extremities.

Attention may here be drawn to the mottled appearance which can be seen in fig. 4 (Plate 3) in those parts of the lens where the crystals have not been formed. These markings, or "globules," as they are termed, evidently represent the stage previous to the actual crystal formation, and must have covered the whole surface of the lens before the crystals themselves were developed. As crystallisation set in they disappeared in the neighbourhood of each crystal, as if the crystals themselves required this material for their formation. The existence and behaviour of these globules are well known in crystallography, as I have since been informed by Professor J. W. Judd, of the Royal College of Science, South Kensington The whole phenomenon is clearly described and illustrated by Professor O. Lehmann in his magnificent work entitled Flüssige Kristalle (Wilhelm Engelmann, Leipsig, 1904). In this conuection Professor Lehmann writes (p. 112), "Das die Globuliten in der Nähe von

Kristallen verschwinden, so dass um diese ein Hof ensteht, erklärt sich vor allen dadurch, dass überkühlte Schmelzen leichter löslich sind als die kristallinisch erstarrte Substanz, in einzelnen Fällen auch durch Strömungen, welche teils durch Dichte-Differenzen, teils, speziell bei Präparaten ohne Deckglas, durch die Oberflächenspannung veranlasst werden."

Although the microphotographs reproduced with this paper indicate in a general way the structure and crystalline nature of the markings, some difficulty was met with in attempting to secure satisfactory photographs with high powers, that is, with magnifications of 150 diameters or more.

The thickness of the crystals and their deposition on a curved surface rendered it impossible to focus the objects in their entirety in a satisfactory manner, so that one had to be contented with the somewhat blurred images when these magnifications were employed.

In order to test whether the markings on the two surfaces of the objective were responsible for any absorption of light passing through the combination, the objective was set up on the V-stand mentioned previously and photographed. On this occasion the background used was white instead of black, and light was only allowed to enter the lens by reflections from this surface. If the markings absorbed light, it was expected that the resulting photograph should give one the picture of the lens with the markings dark instead of light, as was shown in the previous illustrations (Plates 1 and 2). This is exactly what occurred, and both sets of markings were faithfully recorded. Those on the back lens, which covered the larger area, were not, fortunately however, so opaque as those on the front lens, but still a considerable absorption of light was indicated, a quantity quite sufficient to detract from the efficiency of the combination.

I

Inquiry was next made into the probable causes of the deposition of the crystals. The weathering of glass surfaces, on account of the dust, water-vapour, etc. in our atmosphere, has received considerable attention in Germany, and quite a mass of literature now exists on the subject. A very convenient summary of such investigations will be found in Dr. H. Hovestadt's book on Jena glass, and its scientific and industrial applications, an English edition of which appeared in 1902 (Macmillan & Co.).

In chapter x. of this volume the chemical behaviour of glass surfaces is dealt with, and attention is chiefly confined to glasses which are most closely related to the work of the Jena glassmaking laboratory. Since the photo-visual lenses here discussed are made from material supplied from Jena, the information given is therefore specially serviceable on the present occasion. There seems little doubt that it is chiefly the moisture in the atmosphere, and not dust particles, which is responsible for the actions which occur on the surfaces of the lenses here described, for the surfaces attacked are all inside, and although protected to a great extent from dust, they are still liable possibly to action by water-vapour.

It has been shown that the durability of glass depends chiefly on its power of resisting attack by water. Glass being hygroscopic, absorbs moisture to a varying degree, and this moisture enters into chemical combination with its surface. As a result of this absorption, the alkaline components of the glass are gradually set free, and an opportunity is at once afforded for the carbonic acid of the air to combine with the alkalies and form carbonates, which are deposited in the form of crystals.

Dr. E. Zschimmer* made in 1901 an analysis of about 200 pieces of glass with polished plane surfaces, which had been stored, with a view of subsequent testing, in a dry place in Jena for several years, but so protected that they were only imperfectly excluded from air. He found that the behaviour of silicates without lead depended almost entirely on the amount of alkali contained in them. If the amount was as small as 10 per cent., dust particles initiated a deposit which he termed "dusty disintegration"; over 10 per cent., a "homogeneous decomposition" set in which attacked the glass uniformly. When the amount of alkali reached 20 per cent. the deposit became visible to the naked eye, and the greater the proportion of alkali, the coarser the phenomena, i.e. the formation of drops and the crystallisation of carbonates. The action of deliquescent carbonate of potash is to cover the surface with more or less minute drops, while carbonate of soda, being only slightly hygroscopic, covers it with assemblages of crystals.

In the case of the lenses dealt with in this paper, the formation of the single crystals took place on the baryta light flint glass, while the assemblages of crystals occurred on the light silicate

crown.

In order to find out the actual composition of the latter glass, Messrs. Schott and Genossen of Jena were communicated with, and asked whether they would be good enough to supply this information. Needless to say, the requisite data were soon forthcoming, and I should like to take this opportunity of expressing to them my best thanks for so willingly sending me the facts desired. The following is an abstract of their reply, dated 12th April 1907, which they have kindly allowed me to publish :

"We beg to say that the glass in question of our type O 374 constitutes a crown glass holding a high percentage of alkali, we had made at some previous time, of a similar composition as French glass. Owing, however, to its being wanting in permanency, we discontinued later on the manufacture, and hence this glass does not appear on our present catalogue.

2

"Its chemical composition is the following:

B20 3%, K20 17%, Na,O 5%, CaO 3.8%, PbO 1%, As2O3 0'4%, SiO2 69.8%."

With these facts before us, one is led to conclude that the

* Chemiker Zeitung, 1901, 25, No. 69.