D, of the lamp; and on the opposite side an aperture, E, closed with a cork, through which the spirit is poured into the outer chamber. The horizontal bearing-ring, F, fixed to a vertical bar, slides up and down in the handles, and may be adjusted by the thumbscrew to carry the bath at any required height above the lamp. The inner chamber is provided with a brass cover, G. But for the last four years I have restricted the use of this lamp to the purpose of producing dry heat, which it does very well indeed, by simply enclosing it in a thin iron case, having

a lid pierced with numerous holes, each too fine to allow of much flame passing through; enclosed in a box the sixth of the size of the large standing bath, it speedily raises the thermometer to 160°, or even much higher when the experiment is continued long enough. When it is employed for this purpose the bearing-ring is superfluous. Some persons, however, prefer this lamp for the vapour on account of the rapidity with which it yields the heat, and in this case the ring must be retained.

The heat is confined by means of an extremely light,

elastic flanneled hoop, which, though it expands as seen in the engraving, can be rolled up, tied, and packed away in a drawer. The outer covering is made of mackintoshed muslin, and the inner of the finest Welsh flannel, which can be taken out and washed whenever it is thought desirable. The framework consists of a hoop or crinoline of very slender steel bars.

As some of my readers might be disposed to try further

Fig. 6.

季

Fig. 7.

experiments with a view to improving the bath, I will take the liberty of adding a few words as to the failures made, and the obstacles I encountered before completing the apparatus now recommended.

time and money to do so.

It may save waste of

It was thought that a flat surface, with hollow cones rising from the base, would offer the required conditions, and two basins were accordingly constructed on this principle, as shown by the above diagrams (figs. 6 and 7),

in which A A indicate the cones, and the arrows the supposed direction of the flame. Trial, however, did not fulfil the hopes entertained; especially it was found that, the larger the cones, the more rapidly was the evaporating surface diminished when a thin layer of water was used; and when a deeper stratum was employed, too long a time was required to heat it. Besides, with a flat bottom the bath very easily slipped off the bearing-ring, and on two or three occasions was upset. I tried to obviate this by means of a circular lip, and then, as it was found that the latter put out the flame by cutting

ннн

Fig. 8.

off the access of air, prongs were substituted, but they got entangled and bent; besides, they made the bath very difficult to pack. Lastly, an attempt was made to convey the heat by means of small upright pipes sinking through the bottom and rising a little way up in the fluid, as seen in fig. 8, where the dotted line shows the level of the water, just above the mouths of the pipes, but the experiment was not successful, neither did the substitution of solid pegs prove more fortunate. And on the whole, I may safely say that in my hands no mode of applying heat answered so well as that now in use.

With respect to the material employed, I experimented

chiefly on copper, glass and tin. Copper did not answer well at first, but Mr. Walters seems to have quite overcome the difficulty, and is now making baths of this metal. Thin glass, such as is used for well-made mixing jars, certainly enables us to boil water very quickly, but it is unsuited for packing, it is excessively apt to slide off the bearing-ring, and is liable, even when turned out by very good makers, to explode. One jar on which I was experimenting, and which I was assured might be safely exposed to heat, flew into a score of fragments. I have not tried iron; so that perhaps I ought not to speak on the subject, but unless coated with some substance capable of resisting oxidation, such as silver, it would be liable to rust; if thin it would bend very easily, and if too thick would require much longer to heat; so that after numerous trials I have been led to look upon tin as the most suitable material yet employed. Properly coated with nickel, it answers admirably; and if the reader do not mind going to the cost of having the work done efficiently, can be safely recommended as superior to anything else for the body of the bath, being light, strong, and easily kept clean. The expense is no doubt an objection, but I would warn everyone against any economy in this particular. Tin, when unprotected, will not do for a continuance. However carefully dried, it speedily rusts, especially at the junction of the pipes with the lower edge of the basin; and directly the process of oxidation is set up, it becomes simply a question of time how soon the bath will leak. The tubes, Mr. Walters has, so far, been obliged to construct of brass.

It was necessary to determine the amount of spirit

fuel, or methylated spirit, required to the bath boiling for a certain time.

keep the water in For this purpose a

long series of experiments was demanded, and the earliest of these were all carried on with the blast lamp. At first sight it would seem a very simple affair, but the problem proved so complicated that upwards of forty observations were requisite to get at anything like certainty. The first question to be decided was, what quantity of spirit in the outer chamber would produce a blast of sufficient duration to go out at the very same time with the flame in the inner chamber. On this head I could find no information. In Griffin's excellent 'Chemical Handicraft' we are directed to pour into each chamber as much spirit as will fill the cover of the lamp; but although this no doubt answers very well for the ordinary requirements of chemistry, it turned out to be quite inadequate for the bath; and it was only after many failures that I was enabled to determine that, with good methylated spirit, the proportion in the outer chamber must be two and a half to one in the inner Scarcely had these observations been brought to a satisfactory conclusion, before it was discovered that the blast thus obtained was too that it would be absolutely necessary to reduce the quantity of spirit in the outer chamber to eight drachms. When the ring-flame was substituted for the blast lamp, it became necessary to go over many of the experiments again. As, allowing for the cooling of the lamp, each trial occupied about half an hour, and as each stage of

strong for any lamp, and