42 
PHYSICS: J. A. ANDERSON 
Proc. N. a. S. 
existed which was directing its actions. Furthermore, during the time 
in which this remarkable behavior was taking place the hermit crab, 
although at other times reacting quickly to movements outside of the 
aquarium, showed almost no reaction to the stimuli which are ordinarily 
produced by people passing or moving in front of the aquarium. This 
account describes a single series of observations but several other series 
were made in which the behavior was similar. 
It seems probable that the anemones become associated with the hermit 
crabs very early in life, and that possibly they become attached in larval 
life to mollusc shells inhabited by hermits. 
While the anemones are undoubtedly benefited by their association with 
the hermit crabs since the latter lead a wandering life in heavy growths 
of eel grass which harbor many small organisms, and while they may often 
take morsels from decaying animals upon which the hermits feed, the 
writer has never seen the hermits transfer pieces of food to the sea anemones 
as has been described by one investigator. 
In the absence of any experimental work on these animals it cannot 
be proved that the sea anemone affords any protection for the hermit but 
it seems probable that stinging cells of the former would prevent the her- 
mits being eaten by fishes. 
We cannot believe that the hermit crab during its life time has learned 
by experience that it derives any advantage from transplanting the sea 
anemone although we know that crabs in general do profit by experience, 
and yet assuming that the remarkable behavior of the hermit is due to 
instinct, that is, to an "inherited combination of reflexes" which have been 
so brought together by the nervous system that the behavior has become 
fixed and adaptive in the species, it is extremely difficult to conceive how 
it has come about. 
SPECTRA OF EXPLOSIONS 
By J. A. AndkRSON 
Mount Wilson Observatory, Carne;gie Institution of Washington 
Communicated by G. K. Hale, December 2, 1919 
Some experiments with explosions of fine, iron wires by electrical means 
have led to the development of a method of producing a brilliant con- 
tinuous spectrum extending throughout the visible spectrum and as far 
into the ultra-violet as can be recorded by a quartz spectrograph. On 
the continuous background appear as absorption lines practically all the 
iron lines usually found in the arc. For example, in the green region be- 
tween X5150 and X5700, which is unfavorable for reversals in general, 
this spectrum on a scale of 6 A per millimeter shows 128 absorption lines 
due to iron; Burns' table for the same region contains 215 lines, and 
