cele"brated Arctic explorer, show that snow which fell in Sweden 

 contained iron, silica, phosphorus, and cobalt, the quantity of 

 the last named being relatively great. 2. In a communication 

 to Knoivledfje, an F.E.A.S. claims to have found in hailstones 

 which fell at the period of a fine glow green plant cells, spores 

 of fungi, a few diatoms, and some rotifera or their eggs ? 

 3. Mattieu Williams examined snow that fell at Harrow, and 

 found large quantities of black oxide of iron easily separated 

 by the magnet from the remainder of impurities. Some of his 

 experiments leading him to infer the presence of nickel in the 

 iron, which is characteristic of meteoric iron, he considers it 

 to be meteoric. 4. Mr. Eanyard examined this and found it 

 satisfactorily free from terrestrial dust. 5. Of McPherson's 

 analysis we learn that the snow he collected contained 

 hypersthene, pyroxene, magnetic iron, and volcanic glass, all 

 of which were ])resent in the volcanic ashes — (6) whilst M. 

 Eenard states that he finds the dust which fell at Batavia from 

 Java to consist mainly of glassy particles, among which may be 

 distinguished plagioclase, augite, rhombic pyroxene, and 

 magnetite, the silica amounting to 65 per cent. JS'ow, although 

 the silica amounts to over half it has not been noticed in, the- 

 majority of dust samples dragged out of the air by snow, rain, 

 or hail. Again, there is no proof that the snow, &c., might not 

 have brought its impurities from the lower strata of air most 

 likely to be dust laden. 



The absence of the rainband in the spectrum is no proof that 

 the air does not contain water in some form. Thus, although 

 it has been shown by a number that the glow has been existing^ 

 when the spectroscope has shown an absence of watery vapour^ 

 the results of an investigation by F. AY. Cory, of England, and 

 presented by him to the Meteorological Society, show that a 

 rainband as high as 75 per cent, was followed by a slight rain,, 

 and one of 10 per cent, by a heavy fall of snow. They also shew 

 that the spectrum is not affected previously to a snowstorm ex- 

 cept negatively, as the rainband diminished several daysbefore- 

 snow. He suggests that when vapour is transformed into 

 crystals of snow it does not give a rainband. Observations, 

 with the rainband spectroscope are not accepted by all 

 meteorologists. "Eor example," says Hazen, "if it is turned 

 to different parts of the sky and then to a white wall 50 feet 

 or so distant it will be found impossible to tell between the- 

 rainband of the whole sky and that of the 50 feet," and "yet,"' 

 he says in his paper to the Philosophical Society, Washington, 

 " we find it insisted that the instrument must be turned to- 

 different points in the sky." 



I thought at first that this objection was hardly fair, and I 



