846 REPORT— 1898. 



and the ligbt of two or even of one candle can be accurately made. But a material 

 error is less likely to occur where the atmosphere may be foggy, and in a number 

 of routine observations, if the two lights compared are more nearly equal For 

 this reason in technical photometry it is better to make the standard of comparison 

 a light of 10 or IG candles. An actual cluster of so many candles would give a 

 much more constant lijrht than two candles, but its use on a photometer presents 

 obvious and insuperable difhculties. Hence tbe need of a large but compact 

 standard flame. 



After many trials of Argand lamps with wicks and chimneys, the author 

 concluded that the glass chimney was a source of variation, and that if possible 

 an Argand lamp without a chimney must be produced. The result of many triah 

 to produce a lamp of the right kind, and many adjustments, first large, then small, 

 to obtain from sucli a lamp a constant light, and a light of exactly 10 candles, has 

 been the lamp which is now before the Section. 



The burner is supplied with a mixture of air and gaseous pentane from a 

 reservoir carried on a bracket at the top of the lamp. As this mixture falls down the 

 siphon connecting the two, fresh air enters the reservoir, which is provided with 

 cross partitions causing the air to travel backwards and forwards over the surface of 

 the pentane, and to mix with a proportion of pentane vapour, always large, though 

 varying in amount with the external temperature. The variation in the proportion 

 of pentane thus occurring does not aflect the output of light under the other 

 conditions about to be described. A casing round the burner with a conical top 

 steadies the flame, the npper part of which is drawn together into a long brass 

 chimney which cuts oft' the light of this part of the flame. Round the chimney is 

 an outer tube, open below and connected above with a longer tube, which 

 descends and is connected below with the central chamber of the burner. The 

 longer tube is kept cool by having attached to it the bracket carrying the reservoir 

 in which the pentane evaporates, and also a triangle of blackened copper which 

 supports the bracket. Thus an air-current is produced, ascending in the heated 

 and descending in the cooler tube, which issues through the middle of the Argand 

 burner. 



A steady flame of a height between 60 and 70 mm. is thus formed, gi\dng a 

 total light of rather more than 10 candles. By setting the tube which receives 

 the top of the flame at a height of 47 mm., the light shed horizontally is reduced 

 to exactly 10 candles. The total height of the flame can be observed through a 

 small talc window in the side of the chimney, and regulated by means of a tap on 

 the outlet of the reservoir. A variation of a centimetre in the height of the flame, 

 or of a millimetre either way in setting the height of the chimney above the 

 burner, makes no measurable difference in the light emitted. 



As the lamp is tall and its centre of gravity rather high, and as an upset would 

 cause a spilling of pentane which might be dangerous, a firm support is required. 

 A tripod, which for levelling and stability is best, has the disadvantage that, unless 

 the branches ai'e very long, it oft'ers a weak resistance to an upsetting force in three 

 directions. The stand of this lamp has been strengthened in these three directions 

 by being provided with three additional branches, whose screws are turned up so 

 as not to touch the table till the lamp resting on the other three branches has been 

 set upright. The screws of the three supplementary branches, which are made to 

 tui'n very easily, are then turned down in succession till a slight resistance shows 

 that each ia just touching the table. 



A number of comparisons have been made of four of these lamps one with 

 another, and between the lamps and the 1 -candle standard. The results show 

 that all the lamps give the same amount of light, and that this light is exactly ten 

 times that of the 1 -candle standard. 



7. On a Convenient Form of Drying Tube. 

 By A. G. Vernon Harcourt, FM.S. 



A common method of drying gases is to pass them through a wash-bottle 

 containing sulphuric acid, and then through a U-tube filled with fragments of 



