the nalural Production of Saltpetre. 419 



It was an obvious supposition that the rejMoductlon of the 

 nitre was connected witli some change in the state of the at- 

 mosphere : and it is to be remarked, that its first-mentioned ap- 

 pearance took place in frosty weather, and that its reappearance 

 about the 1 Gth of March liad been preceded by a frost of a few 

 days ; whereas during the intervals in which it disappeared and 

 was not again produced, the weather had been mild. The wall 

 was again brushed quite clean on the 16th of March; between 

 which day and the 4th of April a considerable quantity had 

 again collected on the same part ; the process liaving been more 

 rapid during the last four or five days, which were cold. 



On April the r2th the front of this part of the wall was re- 

 newed by scraping, for the purpose of observing what would be 

 the effect of a fresh surface, and before the 25th of April nitre 

 had formed on many parts of this new surface ; but having first 

 increased and then continued stationary for soniC days, it began 

 to diminish about the beginning of May, the state of the air having 

 been very moist during the two or three preceding days. 



About the middle of May 1813 1 selected several distinct 

 parts on which the spontaneous formation of nitre usually takes 

 place, some within and some without the building of the Mu- 

 seiun, and began to make daily observations on tiie phaenoniena 

 of that process, which I noted down at the time: but as a re- 

 gister of this kind v.-ould l)e unnecessarily tedious in its recital, 

 I shall beg leave to give the following general results of those 

 observations ; requesting it may be kept in mind, that i do not 

 presume to lay a greater stress on them than is proportional to 

 the short period of time and the circiunscribed space in which 

 they were made. 



It aj)j)ears then from the observations 1 have hitherto been 

 able to make, that the spontaneous formation of saltpetre is, 

 generally speaking, much more extensive and rapid in winter 

 than in summer, whether it take place on the interior or ex- 

 terior of a building : thus during the period of the cold weather 

 in January 1814, it became visible not only on parts of the walls 

 where I liad never before observed it, but even on the pave- 

 ment of the laboratory. The part of the pavement on which it 

 appeared is adjacent to tlie north wall ; but as the greater part 

 of the pavement is covered by a flooring of wood, it is impossi- 

 ble to say whether or not it took place on other parts also. The 

 stone forming the pavement is the same kind of limestone as 

 that of vvhich the walls are Imilt. 



Again, whereas in sunnner its reproduction is most rapid and 

 extensive in proportion to the degree of light ])resent, the re- 

 verse of this (though not imiversally) takes place in winter. 

 Whercver the saline clilorei-cencc iu qucbtiou occurs, the sur- 

 D d 2 ' face 



