274 
POPULAR SCIENCE REVIEW. 
tion of gadolinite and samarskite, lias always laid great stress on the atomic 
weights. As soon as a substance shows an atomic weight higher than that 
of the other members of the same group (for example, decipin and ytterbin) he 
regards the existence of the new element as established. When, however, 
its atomic weight comes between those of other members of the same series 
(for example, philippin compared with yttria and terbia), experiments were 
continued until the chemist was convinced that the body under investigation 
was not a mere case of a mixture of others. In such cases a study of the 
physical characters was of value. On these grounds the author regards as 
established the existence of ytterbium, decipium, and philippium. Of scan- 
dium, he can say nothing, as he has not met with it. Of mosandrium, he is 
firmly of opinion that it must be struck out of the list of elements. On 
samarium, Lecoq de Boisbaudran published a note, in August, 1879, and in, 
the preceding February, one on another new earth in samarskite. The pro-' 
portions of the latter appear to differ but slightly from those of decipin, or a 
mixture of that oxide with terbia ; and the character of samarium, chiefly 
known by its absorption-bands, leaves the existence of that substance doubt- 
ful. Marignac and Soret have shown erbin to be a mixture of several bodies ; 
and this result Cleve, with the help of Thalen, has confirmed. Cleve goes 
even further in giving- names to two metals which he distinguishes by their 
spectroscopic differences. The extreme Ted line noticed by Soret, he refers 
to thulium, and the red and green lines (X = 640 and 536) he terms holmium. 
Samarskite contains but little of the old erbia ; the spectrum of its solution 
shows it but faintly, especially the green and red absorption-bands (X = 523 
and 488) ; the indigo fine, on the other band (X = 462), is stronger. After 
removing the didymium, decipium, and terbium compounds, Delaf ontaine sub- 
mitted their formiates to fractional recrystallization : the first products gave 
49-47 per cent of base ; they were purified and again recrystallized, whereby 
a product rich in terbium was obtained, and in the mother liquors another 
rich in yttrium. The residue, rendered incandescent, and then converted 
into nitrates, showed the indigo band 452 strongly ; the others weaker. 
This nitrate, whether dissolved or in crystals, was completely colourless, and 
had an equivalent = 98. By fractional decomposition, and by ignition, it 
was separated into six products (A to F) : A had an equivalent = 102, the 
formiate had a light rose colour, and the absorptive spectrum showed the line 
of impure erbia very strongly. B to E was more strongly coloured, the nitrate 
colourless, the spectrum showed only the band in the red 640, in the green 
536, was small ; and in the indigo, 448-445, broader ; the line of pure erbia 
was reduced to a minimum. The earths, in B to E, cannot be separated into 
terbia and yttria in the usual way. The product F, the last we have to con- 
sider, gives a very weak spectrum ; and its colour, as well as its equivalent, 
was less than that of the substance in the original material. The author has 
given the name philippin to the yellow oxide differing from terbia, which 
forms the chief constituent of the oxide above referred to. Its properties 
accord with those of the X earth of Soret, and the holmin of Cleve, so that 
the latter name should fall into disuse. Philippin may be regarded as con- 
sisting of two oxides, one giving the indigo line 448-445, and the other the 
band 640 and 536 ; the author, however, has met with no facts which favour 
this view. The so-called euxenite of North Carolina contains even less of earths 
