I8a0.] 



THE CIVIL ENGINEER AND ARCHITECT'S JOURNAL. 



299 



his recent work on tlie Skerryvore Liglithouse.^ An octagonal 

 frame, carrying eight lenses, was made to revolve with various 

 degrees of rapidity, and the light was ohserxed at a distance of 

 ll miles. It was then found, that as the rate of revolution was 

 accelerated, the apparent brightness and volume of the flashes 

 diminished; until when a velocity of eight or ten flashes in a second 

 was obtained — the light became almost invisible. Mr. Stevenson 

 correctly explains this result by supposing, that when the lenses 

 revolved rapidly the light had not sufiicient time to produce its 

 full effect on the eye. While these experiments sufficiently prove 

 the impossibility of obtaining the result Captain Mall had in 

 contemplation, yet in the absence of definite information regarding 

 the connection which subsists between the apparent brightness of 

 an object, and the time during which its light has acted on the eye, 

 it is obviously impossible, without actual trial, to assign the limit 

 to the velocity of a revolving light. This information is supplied 

 by the author's researches on the time required for the production 

 of luminous impressions on the eye, published in the 'Transactions 

 of the Koyal Society of Edinburgh' for 1819. His experiments 

 were conducted by means of an apparatus, consisting of two screens, 

 each with circular apertures an inch in diameter, to which are 

 fitted pieces of ground glass. Tlie apertures are illuminated by 

 gas burners, which admit of having their distances from the screens 

 varied at pleasure, by sliding their supports along the plank in a 

 groove cut in it for that purpose. The brightness of the apertures 

 in the screens is observed by means of a rectangular prism of glass 

 placed half-way between them, with its faces inclined at angles 

 of 45^ to the line joining their centres. By this means the aper- 

 tures are seen in apparent contact, and their relative brightness 

 can thus be compared «itb great nicety. A disc is made to re- 

 volve on the axes in front of one of the screens, having a sector of 

 a known angle cut in its circumference; and in this manner the 

 aperture is seen for a short interval of time at each revolution of 

 the disc. The time during which the light acts on the eye is easily 

 ascertained, by knowing the velocity of the disc, and the ratio 

 the arc of the sector bears to its whole circumference. 



Before causing the disc to revolve, the apertures in the screens 

 are made equally bright by varying the distance of the light from 

 the screen. When the disc is then made to revolve, the apparent 

 brightness of the aperture behind it is instantly diminished; and 

 the equality of the brightness of the apertures in the screens is 

 restored by increasing the distance of the light from the screen. 

 The ratio of the brightness of the impression produced by the light 

 during the revolution of the disc to the brightness of the im]iression 

 when seen by uninterrupted vision, is that of the squares of the dis- 

 tances of the light from the aperture before and during the revolu- 

 tion of the disc. By means of this apparatus, it was found. — 1. That 

 when light of a given intensity acts upon the eye for a short space 

 of time, the brightness of the luminous impression on the retina 

 is sensibly proportioned to the time during which the light con- 

 tinues to act. Thus, the intensity of an impression made in -01 

 seconds is almost exactly Vijt'i of the brightness of the light when 

 seen by uninterrupted vision; and the intensity of the impression 

 is exactly doubled in 'OZ seconds. — 9. Lights of every degree of 

 intensity produce their impression on the eye with equal rapidity. 

 — 3. The time required to produce a complete impression is about 

 one-tenth of a second. 



The conclusion to be drawn from these experiments seems to be, 

 that in any proposed revolving light, the velocity of rotation ought 

 to be regulated so that the duration of each flash must at the very 

 least exceed J^th of a second. This velocity can be easily calcu- 

 lated, for the arc of the horizon, included by the brightest portion 

 of the flash, is equal to the minimum divergence of the rays; and 

 this, again, is equal to the angle which the horizontal diameter of 

 the flame subtends at the edge of the aperture of the lens or re- 

 flector.t 



If, then, t = the time of a complete revolution, t' = the dura- 

 tion of the flash, and a = the divergence of the rays, t == ■ 



a 



Or, if we take t' = -inth of a second as the shortest allowable 

 diameter of the flash, we shall have a< = 72. Thus, in the case 

 of the lens of Fresnel's revolving light of the first order, the 

 minimum divergence is i° 44'; and the greatest velocity of rota- 

 tion that could be employed without necessarily diminishing the 

 apparent brightness of the flashes would be 7-6, or nearly 8 seconds. 

 In like manner the greatest admissible velocity for a parabolic 

 reflector, whose focal length is 4 inches, and its greatest double 



• See Slevenson's 'AccouDt of llie Sk. rryvnre Lipiithouse,' p. 31."?. 

 t See Steveofun's *Accouut of the Skerryvore I.igh;h>juse,* p. 213, 2i7. 



ordinate 21 inches, illuminated by a flame one inch in diameter, is 

 one revolution in nearly 7 seconds. 



In stating these cases, it is not of course assumed that so great 

 velocities would be found suitable in practice. All that can be 

 inferred from the experiments with certainty is this, that any pro- 

 posed arrangement which should employ Fresnel's great lens with 

 a velocity exceeding one revolution in 8 seconds, would necessa- 

 rily be disadvantageous; or more generally, that in every light- 

 house arrangement, care must be taken that the flashes of light 

 have time to act on the eye for more than one-tenth of a second. 



COOLING THE ATMOSPHERE OF BOOMS IN TBOPICAL CLIMATES. 



On a Method of Cooling the Atmosphere of Rooms in a Tropical 

 Climate. By Professor C. Piazzi Smyth, of the Edinburgh Ob- 

 servatory. 



The difficulty of efl^ecting this object is so great, even in cooler 

 countries, that while the apartment of a sick patient during winter 

 is preserved carefully and easily, by means of a fire, at any desired 

 temperature, if this be much exceeded during a few days in sum- 

 mer by the atmos]i]iere, although the patient may visibly suffer 

 from the he.it, still the case seems to be thought so hopeless that 

 the physician and friends are generally content merely to lament 

 the untoward warmth of the weather; or, perhaps, in a few cases, 

 to try to counteract some of the minor consequences of that pre- 

 judicial cause. 



If the problem now proposed is to be solved in all its entirety, 

 it must be stated thus: — 



In a country where the thermometer stands at or above 80° 

 Fahrenheit all through the 24 hours, both summer and winter, and 

 where there can be no coolness in springs, wells, rivers, or the 

 night air; and where t/ic atmosphere is saturated u-ilh moisture, so 

 that no cold can be produced by evaporation, — to lower the tem- 

 perature of the air in rooms; to keep up a constant supply of pure 

 cold air; and to remove that which has been warmed or otherwise 

 vitiated. 



To meet such a case, the present Indian methods are utterly in- 

 adequate, for the punkah, in its various forms, merely serves to agitate 

 the air, and does not cool or purify it in the slightest degi'ee. The 

 wet mats which in some places are hung before windows, and 

 being blown through, naturally or artificially, cool the transmitted 

 air, would be inapplicable in the case now before us, where the 

 air is saturated with moisture. And even where they can be em- 

 ployed, their use is objectionable on the ground of their adding so 

 much moisture to air already overloaded with it; for it cannot be 

 too strongly borne in mind, that in warm countries, though the 

 air may often feel dry to the human frame, that still, on account 

 of the air's capacity for moisture increasing with the temperature, 

 there may be a far larger amount of watery vapour ))resent than 

 even in a Scottish mist. 



And these are all the methods which have yet been brought for- 

 ward for the relief of suffering humanity; for the bane to be 

 removed is the too high temperature of the food of the lungs and 

 the skin. The use of cold meats and iced drinks for the stomach 

 must be regarded as a forlorn hope, and a mistaken idea. 



A complete remedy would, however, seem to be presented in 

 the property of air to increase in temperature on compi-ession, and 

 diminish on expansion — a fact strangely overlooked for this pur- 

 pose, seeing how often workmen are stumbling upon it, while 

 every book on Hydraulics contains an account of the Schemnitz 

 machine, where air rushing out from great compression freezes the 

 drops of w ater that issue with it; and every work on Natural Phi- 

 losophy describes the syringes in which, by the sudden compression 

 of air, tinder is ignited. 



AVhen cold air is to be produced in this way, it is evident that 

 as the quantity of heat continually decreases for each succeeding 

 atmospliere of pressure, it is desirable so to arrange the machine 

 as to compress the gi-eatest possible quantity of air the least degree 

 necessary to produce the required temperature, than to obtain 

 exceeding cold by compressing violently a small quantity of air, 

 and diluting- that afterwards with a larger quantity of common 

 air. 



After some consideration,* a compression of one-fourth an at- 



* If maehinery could be esecuted perfectly, a high degree of compression mi^ht be 

 exerted, and the compressed and cooled air might be made to react in its expansion of 

 the bacit of the pi-ton wh]ch is cor-ipressing the air in front of it; thus savins all the 

 power exerted but that lost io friction, diminished bulk of the cooled air, and imper- 

 fect expansion and leakage ; but on submitting the matter to calculation, there did not 

 seem to be any possibility of adopting th« method with advantage in the present mode- 

 rate degree of perfsctioD of pneumatic engines. 



40* 



