depth than the outer surface of the scales; they are tubular, the whole plant being without septa forms a single individual of apparently indefinite extent. The spores are variously shaped at different stages, ovate and kidney being the commonest forms. They are very minute, and require a power of 450 to observe them well. The cilia are two in number, a longer and a shorter one, and are situated at the long axis of the spore. They are difficult to observe, and always disappear in permanently mounted preparations, although the spores themselves remain unaltered in all other respects. When the fungus is stained with logwood or picric acid, excellent permanent preparations can be got. It has been stated that the fungus dies with the fish. I have not found this to be the case; on the contrary, all my observations have been made from dead fish. Some of the specimens sent me from Carlisle by Mr Dunne were mis-sent to Aberdeen, and returned to me on the seventh day after the death of the fish, and yet I have scores of permanent preparations from these specimens, which show distinctly the characteristic form of saprolegnia ferax.

I have also found the fungus perfectly identical in all the specimens I have examined, which consist of salmon, sea trout, and river trout from the Eden, and salmon and grayling from the Nith.

It has also been said that a salt solution destroys the fungus, "which melts in the solution like sugar in water." On the contrary, salt and water is an excellent preservative of saprolegnia; masses of it before me as I write have been in a salt solution for two months, and it remains unaltered. Further, the salmon captured in the Nith, which is believed to have gone to the sea in order to get rid of the fungus, had the fungus growing vigorously on several parts of its body. The fungus must either have instantly attacked the fish on its return to the river, or not have been destroyed during its stay in the salt water.

Regarding the cause of the disease, I can offer no opinion further than that some functional condition of the fish seems necessary for the propagation of the fungus. The germs of saprolegnia ferax must exist at all times, and in many places; and if so, there must be a reason why fish are not constantly affected with the fungus and in every river. I am persuaded that the condition of the fish

is in some way either suitable or unsuitable for the propagation and growth of the fungus. Whether this arises from too high or too low condition, I am quite unable to say; but I may remark that while some of the fish examined were in the kelt stage, others were in a condition perfectly fit for food.

Saprolegnia ferax parasitic on the Salmon. a. Barren filament. b. Prolific filament. c. Prolific filament, more highly magnified.

4. On some New Bases of the Leucoline Series, Part I.
By G. Carr Robinson.

5. On the Crystallisation of Isomorphous Salts.
By G. Carr Robinson.

It is generally stated that isomorphous salts are capable of crystallizing together in any proportions, or that the isomorphous elements which enter into them are capable of replacing one another in any proportion; e.g., potash alumina alum and potash iron alum can crystallise together in all proportions.

Hauer states that mixed crystals of alumina and chrome alum grow in a solution of ammonia iron alum; and, again, he states † that the more soluble isomorphous salt completely hinders the solution of the less soluble, so that if solutions of common alum, chrome alum, and iron alum be mixed, precipitation of the less soluble alums will occur.

The present paper is the result of some experiments made with the four alums-potash alumina, ammonia alumina, potash chrome, and ammonia iron.

The solutions of the two first, potash alumina and ammonia alumina, were obtained by saturating water at 100° C. with the salts, and pouring off the mother-liquor from the crystals that deposited on cooling.

The chrome alum solution was obtained by saturating water with the alum at a temperature carefully kept below that at which the green modification of chrome alum is produced.

Whilst the solution of iron alum was made by saturating water, acidulated with sulphuric acid, with the alum, at about 50° C.

The following experiments were then made and observed :— a. When a crystal of potash chrome alum is placed in a solution of potash alumina alum exposed to the air, the chromium is turned out by the alumina, the interior of the crystal becomes granular, whilst a clear shell of alumina alum grows over it, the faces of which are finely striated.

b. When potash chrome alum meal is added to a solution of potash alumina alum in a well-closed bottle, and kept at very nearly a constant temperature, the chrome alum dissolves, but no replacement takes place.

When a crystal of potash chrome alum is placed in a solution of ammonia alumina alum exposed to the air, the ammonia alumina alum grows on it, there being only very slight replacement.

d. Potash chrome alum meal added to solution of ammonia alumina alum in well-closed bottle, and kept at very nearly a constant temperature, no change takes place.

e. Ammonia iron alum grows on crystals of potash alumina alum. f. Ammonia iron alum grows on crystals of potash chrome alum.

* "Sitzungsberichte," Imperial Academy of Sciences, Vienna, 1860, xxxix. +"Sitzungsberichte," Imperial Academy of Sciences, Vienna, 1866, liii. VOL. IX.

5 D

g. When a crystal of ammonia iron alum is placed in a solution of ammonia alumina alum, replacement goes on very slowly, rust at the same time being thrown down.

h. Potash chrome alum grows on crystals of potash alumina alum. i. If a crystal of ammonia iron alum be placed in a solution of potash chrome alum, the iron alum is turned out, whilst a skeleton in chrome alum of the original crystal is left; if this skeleton be placed in a solution of ammonia alumina alum, the latter alum grows on it, completing the form of the original crystal of ammonia iron alum.

6. On a New Method for the Separation of Yttrium and Erbium from Cerium, Lanthanum, and Didymium. Part I. By J. Gibson, Ph.D., and R. M. Morrison, D.Sc.

The method for the separation of these two groups hitherto in use was first proposed by Berzelius, and has been followed by almost all chemists who have investigated these earths. For the details of this method we must refer to his "Handbuch;" but we may briefly state that it depends on the relative solubilities of the double sulphates of these metals with potassium in a saturated solution of potassic sulphate. The yttrium and erbium double sulphates are said to be easily and completely soluble, while the double sulphates of cerium, lanthanum, and didymium are said to be perfectly insoluble. Wishing to prepare pure salts of yttrium and erbium, in order if possible to obtain the metals and to determine their specific heats, we tried this method, and found that, although it is a good rough method, the separation is by no means complete. We repeated the separation six times, but never obtained the earths pure, the spectroscope always showing the characteristic absorption-spectrum of didymium, provided we examined a sufficiently thick layer of a saturated nitric acid solution. The incompleteness of this method is indeed acknowledged by Bahr and Bunsen in their well-known paper on these metals.* The test given for the presence or absence of these two groups by these chemists was the presence or absence of the absorption-spectra of didymium and erbium respectively. We found, however, that

* Ann. d. Chem. u. Phar. cxxxvii.

not only are the double sulphates of the didymium group somewhat soluble in a saturated solution of potassic sulphate, but that some erbium, if not also some yttrium, is precipitated. Repetition of this process fails to remove the last traces of the didymium group.

After trying various modifications of this method, some of which gave better results, notably boiling the double sulphates with the saturated potassic-sulphate solution and filtering hot, we determined to look for another method, none of these variations being sufficiently good.

We have obtained better results by the following method. After having extracted as much as possible of the didymium group by the old method, the earths are dissolved in nitric acid, and to this solution a large excess of a solution of carbonate of ammonia is added, and the whole allowed to digest for a day in a closed flask, the precipitate being frequently shaken up. The liquid is then filtered and the undissolved residue washed. On acidifying the filtrate with hydrochloric acid, and adding oxalic acid, a precipitate is obtained which, on ignition, yields the oxides of yttrium and erbium, free from, or only containing the merest trace of, lanthanum or didymium, which may be removed by a repetition of the process. Any cerium originally present goes into solution, and must be removed by boiling a solution of the sulphates with carbonate of magnesia. The undissolved residue of the carbonate of ammonia solution still contains yttrium and erbium carbonates, but is much richer in didymium, and must be treated with a fresh portion of carbonate of ammonia solution.

It is essential that the ammonium carbonate solution be neither too concentrated or too dilute, as in the first case some lanthanum and didymium dissolves, and in the latter the carbonates of yttrium and erbium, after dissolving, crystallize out as double salts, leaving almost nothing in solution. The strength we found most suitable was about half saturated.