lxvili PROCEEDINGS. 
years since, on an unsuspected corner of the reagent shelf, an organic 
compound a@-nitroso g-naphthol was found to instantly and completely 
separate them from each other. Such being the case, we may some day 
expect research to be rewarded and the mysterious doors to be opened. 
I must now, however, in the development of my plan, state the 
methods most in vogue for separating any rare earth from a mixture of 
them. Suppose we have a mixture of all or nearly all of these earths, 
and this is the state in which we usually get them from the minerals 
containing them, and that we have, say 5 litres, insolution. We remove 
10c.c. and precipitate all the earths by means of a standard solution of 
ammonia, noting carefully the exact amount required. From this can 
readily be calculated the amount neceasary to precipitate the 5 litres. 
This being determined we take enough of a somewhat weak solution of 
ammonia to precipitate one-tenth of the earths and add it as rapidly as 
possible, with violent agitation, that it may be brought in contact with 
as large a portion of the solution as possible at once. This precipitates 
the most basic portion to a large extent. The mixture is allowed to 
settle and the supernatant liquid is drawn off; after which the precipi- 
tate is carefully washed and the washings are added to the liquid. This 
is again treated with ammonia, another tenth being thrown down. This 
process is repeated till the entire amount is precipitated. The first two. 
or three precipitates are then united, then the next two or three and so. 
on, and each group is again treated in a similar manner, till after some 
hundreds of repetitions there collects at one end of the series a consider- 
ably basic, and at the other a considerably acid, hydroxide, The various. 
precipitations are checked by atomic weight determinations, and when an 
hydroxide is obtained in which the entire ten precipitates yield identi- 
cally the same atomic weights, it is considered as an elemental earth, 
the argument being that no two elements will be at all likely to possess 
the same basic qualities. This will probably give you an idea of the 
time expended and the difficulty experienced in working in this field. 
Ammonia is by no means the only reagent so employed, but every one 
likely to produce different qualities of precipitates is tried. 
Here, then, we have a group of elements whose compounds act. 
differently toward chemical reagents from all other bodies. They 
resemble the alkalies and alkaline earths, 2. e., the Li. and Gl. families in 
their action toward H,S, and the B. and Fe. families in their action 
towards (NH,),S. They resemble the last two and all the other families. 
except the Li, and GI. families in their action toward NH,OH. They 
