In order to remove all doubts concerning the amount of expansion of petroleum, to prove that it does not expand more than whisky, and lessthan alcohol and most of the acids and oils, as stated in my communication to the Scientific American, page 38, current volume—I give here some of the data on which my statement was founded ; and will exhibit only a few of a great number of determinations which I have made to settle this question, selecting those which recommend themselves by simplicity, because of the ro nd numbers obtained First Metliod by Means of the Specific Gravity Bottle. A small bottle, with ground-glass stopper, made so as to contain, when entirely full, exactly 50 grammes of pure distilled water at 65° Fah., was filled with heavy kerosene, the product of the last stages of distillation, marking 30° on Baumes hygrometer; it contained at 32° Fah. exactly 44 grammes of the oil. When heated to 212°, a certain quantity of oil did overflow, and after cleaning and cooling [the weight of a hot object cannot accurately be determined on a sensitive balance, because of the air currents generated ; this as a hint to young chemists] it was found to contain 4115 grammes, proving an expansion of 2'85 grammes, or 6'5 per cent of the whole. As, however, the glass of the vessel expands, according to Regnault one 290th of its volume, this fraction of the 44 grammes has to be added for correction; it is nearly 0'15 grammes, which makes ihe expansion of the oil from 32° to 212° Fah. , equal to 2'85 + 015, or 3 grammes, which is one 14-7th part of 44 grammes, and an expansion of 6'8 per cent., or 0'068. Other determinations with the same oil gave sometimes 0'069, 0'070, and 0'071. Common kerosene of 49° Baume was placed in the specific gravity bottle, and one of the samples gave, at 65°, exactly 40 grammes ; heated to 125° it gave, after correction for glass expansion, 1 gramme less, being 0'025 for 60°, consequently 0'075 for 180°. When cooled to 35° it gave a contraction in bulk of 0'048 gramme, or 0'012th part of 40 grammes, corresponding to an expansion of 0'012 for 30°, or 0'072 for 180°. When heated from 120° to 180°, the expansion was found to give a co-efficient of nearly 0'079. On these facts I founded my statement referred to, that the rate of expansion is less between 32° and 60°, and more at about 180° than the mean expansion, which is 0'076. Light gasoline ot about 90 Baume was experimented upon, one sample gave for contents of spec. gr. bottle at 30° Fah., 33 43 grammes. and at 60° exactly 32 grammes. This- gave an expansion of 0'43 grammes for 30° Fah. of heat. or on". 74th part of the whole, which Would give for 180° a little more than one twelfth, or 0 083—a rate of expansion only slightly larger than ether and turpentine, equal to most animal oils, but considerably smaller than alcohol, nitric acid, , olive, and linseed oil. In crude petroleum the expansion was foundr always between 7 and 8 per cent, and in proportion as they were heavy or light, it was nearer to the first or to the second of these numbers. Second Method by .Means of the Hydrometer. When placing a thermometer and hydrometer in kerosene of 40° Baume, at 65° Fah. temperature, and heating it to 125°, the hydrometer will sink and indicate 46° ; as now 40° Baume corresponds with a specific gravity of 0'83, and 46° Baume with 0'819, it indicates an expansion of 0'83—0'819, or 0021, which is the 40th or 0'025th part of 0'83, this amount for 60° gives 0'075 per 180°, the same as found above It will be found, in general, that for every ten degrees increase of the thermometer the hydrometer sinks one degree lower, and vice versa. For the lighter oils, a little above nine degrees Fah. will correspond with one degree difference in the hydrometer, and for the heavier oils 10'5° to 11° Fah. of heat will be required to make this difference, but in general ten degrees heat for one degree gravity is near enough for practical purposes ; and, in fact, this is so well known that it is depended upon by experts as a necessary correction in determining the quality of different grades of oil. As 50° and 60° Baume, respectively, correspond with a specific gravity of 0'785 and 0'709, the difference of these last numbers, 0'076, correspond with 18° of Baume's scale, which, again, correspond with the expansion for 180° heat. Every degree of Baume's scale corresponds thus with 0076 divided by 18, or 0'00042, nearly, for the corresponding difference in specific gravity. :1 hird Metlwd by Means of the Thermometer alone. When taking a correctly graduated alcoholic thermometer, breaking the top open, heating the bulb so as completely to remove the alcohol, and then filling it with petroleum to such an extent as to make the freezing point of water 32° Fah, on the scale, to correspond with the surface of the petroleum in the tube when cooled to 32°, then placing this thermometer in hot water of 122°, as indicated by another thermometer, then the petroleum thermometer will only indicate about 100° on the scale ; as the scale was constructed for the alcohol, its degrees are as much too large for the petroleum degrees as the expansion of alcohol exceeds that of petroleum ; in this case it is found that 122—32 or 90 parts of alcohol correspond with 100—32 or 68 parts of petroleum ; these numbers—90 and 68—are nearly in the same ratio as 0100 and 0'076, the numbers expressing the ratio of expansion of alcohol and petroleum—another verification of the statements in the table published on page 38, already referred to. When we consider the simplicity and reliability of all these methods, by which the rate of this expansion may be determined, and the perfectly accurate manner in which they cor roborate one another, it is indeed surprising that M. Deville, before the French Academy, dwells so largely on the “ very great expansion in bulk which mineral oils undergo by increase of temperature,” and that when “barreled during the cold season it will expand largely with the first appearance of hot weather, and burst the vessels, on the same principle that ice ruptures our hydrants.” [See Scientific American, page 376). That M. Deville does not communicate the ratio of this, according to him, so extraordinary expansion, is not truly scientific, and makes his whole statement unreliable. I adhere to my opinion expressed before, that the cause of leakage of petroleum barrels by heat, is the elongation of the iron of which the hoops are made, which makes the staves loose ; besides this, the staves will contract from the same cause, which increases the leakage ; add to this the extreme penetrating power and volatility of the lighter portions, chy-mogene, gasoline, etc., which is so largely increased by any rise in temperature, and we have a perfectly satisfactory explanation of the increase of danger in hot weather. Why Coffee is a Stimulant. The changes which heat effects in the elements contained in the green coffee berry have been little studied; we merely know, from the researches of MM. Baitron and Fremy, on the one hand, and of M. Payen on the other, that the brown bitter substance and the aromatic principle are produced by the decomposition of that part of the coffee bean which is soluble in water, and that a large part of the caffeine disappears during the roasting. It is said that this (caffeine) is carried away with the volatile products generated in the operation. By roasting coffee in an apparatus which allows of the recovery of all the volatile products, I have ascertained that if caffeine be carried away with the volatile products, it can only be in such small quantity as is not appreciable by weight, and cannot explain the considerable loss which takes place during roasting carefully performed. The loss is experimentally found to equal nearly one-half of the caff. ine originally existing in the coffee. I have succeeded in demonstrating that the lost caffeine has been transformed into a volatile base— methylamine, or metbylammonia (C4H5N), which was discovered by M. Wurtz. The following are the facts which prove the change of caffeine into methylamine during coffee-roasting : If pure caffeine be submitted to the action of heat, and the vapor be carried through a tube heated to about 300° Cent. (about the heat which is necessary for roasting), and filled with fragments of pumice-stone, which delay the passage of the vaporized matters, only a feeble decomposition occurs; the greater part remains unchanged, and the little that is decomposed gives no characteristic product except cyanogen. This experiment tends to prove that it is not the caffeine which furnishes the volatile alkaloid existing in roasted coffee. But a very different result is obtained if, instead of acting on free caffeine, we experiment on caffeine in analogous circumstances to those in which it exists in green coffee. M. Payen has, in fact, shown that caffeine exists in that berry in the form of the tannate, i, e., a combination of caffeine with a tannin peculiar to coffee. On submitting to the action of heat the tannate of caffeine which has been prepared with tannin of gall-nuts, we obtain, as with green coffee, methylamine : this compound behaves, under the influence of a temperature of about 300° Cent. , in a manner similar to the tannate of caffeine first isolated by M. Payen. The whole of the methylamine produced during the roasting of coffee is not found in the solid residue; a certain proportion escapes with the volatile matters. It is easy to extract the alkaloid from roasted coffee by distilling the exttact of coffee, made with cold water, with a weak base, such l),s lime. The addition of this alkali to an- infusion of coffee immediately liberates the methylamine, the special ammoniacal odor of which is readily pe^ ceptible,—MM. I'ersonne. Advertising Made Easy. At a recent meeting of the “ Society of Practical Engineers,” one of the Society's M. D.'s read an elaborate paper on water meters, and closed with an eloquent description of a meter which the speaker had himself invented. He believed that he had made the most accurate, the most simple, the most durable, and the cheapest water meter in the world' and he invited for it the closest scrutiny and investigation of all concerned, This apparatus, he said, could be furnished at two dollars apiece. This is certainly cheap enough; we had no idea that an accurate, simple, durable water meter could be had for so small a sum of money. Besides, we are pleased to notice the liberal conduct of the learned society, in permitting the discoverer' to eulogize the merits of the meter in the proceedings of the evening, The Secretary of the Society will please send us its advertising terms. The Snorer's Companion is the name of a newly patented device to be attached to the backs of church pews, forming a comfortable head rest, and enabling the owner to sleep through the dullest sermon with the greatest satisfaction. © 1869 SCIENTIFIC AMERICAN, INC.