_ APRIL 7, 1923] 


























animals so that, as is well known, the head of Adamsia 
hung downwards and its tentacles, brushlike, were 
carried over the surface of the sand when the hermit- 
crab travelled from place to place. The first two 
pairs of the long, slender walking legs of the hermit- 
crab were directed backwards in a manner which 
‘suggested protection of the anemone, but this appear- 
ance was misleading, as it was soon found that their 
function, in addition to that of locomotion, was to 
‘steal the food collected by the anemone. This was 
effected most cleverly by an underneath upwards 
sweep of the leg, the terminal portion of which passed 
through the tentacles of the anemone and carried any 
: found therein swiftly to the mouth of the hermit- 
crab. It is interesting to note that these limbs seem 
ially adapted to this purpose. The part men- 
tioned (the dactyl) in this species (E. prideauxii) is 
long and very slender, and its inner or concave side 
is beset with a row of many long fine hairs projecting 
inwards like the bristles of a brush, thus forming a 
very effective instrument for sweeping out the mouth 
of the Adamsia. At times, also, the claws were 
doubled under the hermit-crab’s body and seized the 
food which had been secured by*the anemone. At 
first food was supplied for the joint use of the animals. 
Later on I experimented and tried to feed the anemone 
e, but in this I never succeeded, as although the 
ermit-crab could not see the food, it was so instantly 
detected and swiftly swept away, as described, that 
one wondered how the anemone ever got sufficient 
for its own needs. Whether some sensory hairs on 
the dactylopodite had anything to do with detection 
I cannot say. 
_ My observations seem to show that, though both 
animals benefit, the advantages of the partnership in 
this particular case are very largely on the side of the 
hermit-crab, which, in addition to being supplied 
with food, may possibly derive some benefit from the 
Adamsia’s power when irritated, of firing a broadside 
of stinging threads through the numerous portholes 
‘in its sides. So far as I can see at present, the only 
profit to the Adamsia is that of being carried from 
place to ba and thus afforded a better chance of 
‘securing food, for which, as has often been pointed 
out, the downward direction of the mouth and 
tentacles is most favourable. The anemone may, of 
course, derive other advantages which are less obvious, 
and the parallel case (to which Dr. Orton has directed 
attention) of the little tropical crab, Melia tessellata, 
which carries in each claw a living sea anemone and 
uses it as a weapon and also (like Adamsia) as a 
collector of food, suggests the possibility. 
On the face of it, Adamsia and the little anemones 
first mentioned seem to be the willing slaves of the 
_ hermit-crabs, for P. H. Gosse’s observation, in 1859, 
of how E. prideauxii with its claws transferred the 
_ Adamsia from its old shell to a new one (‘‘ A Year at 
the Shore,’’ pp. 241-247), which was later confirmed 
; wy Col. Stuart Wortley (Ann. and Mag. Nat. Hist., 
1863, p. 388), seems to show that the hermit-crab is 
the keenly interested active agent in arranging matters 
so advantageously for itself. With the common 
hermit-crab (E. bernhardus) and Sagartia parasitica, 
however, matters are reversed. Here the anemone 
evidently takes the initiative, and except perhaps by 
the camouflage, etc., which is afforded by its riding 
on the whelk-shell occupied by the hermit-crab, the 
latter appears to derive no benefit. The position 
assumed by the anemone is unfavourable to the 
hermit-crab’s sharing in its captures ; moreover, the 
walking legs of the hermit-crab are not adapted to 
securing a portion, the concave side of the dactyl of 
E. bernhavdus being smooth and practically free from 
NO. 2788, voL. 111] 
NATURE 
_. Matters had thus been arranged between the two + hairs, whilst the limb is otherwise unsuitable for the 
= SS 
a er 
465 

purpose. It seems as though S. parasitica has taken 
a hint from Adamsia and improved upon it. 
ARNOLD T. WarTSON. 
Southwold, Tapton Crescent Road, 
Sheffield, March 15. 

Paradoxical Rainfall Data. 
Pror. McAnig, in Narure of March 17, p. 362, 
directs attention to the apparently paradoxical fact 
that the wettest month observed in 37 years at 
Blue Hill observatory fell in June, the month with 
the lowest rainfall average, whereas the driest month 
fell in March, the month with the highest average. 
The coincidence is a curious one, but less improbable 
than might at first sight appear, since the monthly 
rainfall is at many stations extremely variable. 
Some idea of its extreme variability may be gathered 
from the following table, showing the distribution in 
half-inch intervals of the rainfall at Rothamsted for 
7° years from March 1853 to February 1923. 

































Pe ar eile pels 5 
/S/2/EZISISISIEIBISlala 
E/BISISB/SI/2/E/2/8/2] 8] 8 
Pe a bl poh lca Wem) =: a/S 1/3] 
o a Zz|A 
° l Lae! = 
PO Fae a (ah ae ll fe el ( el ‘a 
5 | Dead = 
9p Spe BM ef Ou Ss ae OEE 6 
ro — —— = 
124 | 19 | 13 | 12/18 }10 | rr | 13 | 13 fd ie 9 9 
iS — — — 
74 | 10 | 16 | 14 | 12 | 13 7 7 del ei le 4] 14 
20 — —— _ 
10 8 | 10 | ro | 12 64 | 11 8 | 17 5 | 18 7 
| 9 4 7) 12 | 8 |r0ob] 5 | 13 gy eX ax 6 
3:0 |———}|—_|—_ S| | | 
7 S| 4]. 3) Sh Sl Sa) es One Texe 
3°5 = aes i (Se eee 
9 6 4 3], °2 5 2 7 2 5 + 3 
4°0 |-—— —_ 
4 Sh ess ee eee ak (Oe ek 
#5 —— 
2 2 I Sab | os 5 2 I 3 3 I 
5:0 |- Be) 1d) fed) Ae ee 
4 2 Fy Me: Oh ae ay et 
55 <a 
1 2 I I 4 2 
60 = — 
2 I 2 
65 — 
| I 3 2 
7:0 |- er == a 
I 
75 Fri 
8-0 _ 
} I I 
85 I | Se ee 
70 | 70 | 70 | 70 | 70 | 70 | 70 | 70 | 70 | 70 | 70 | 70 
! 







The seasonal effect appears to be more strongly 
marked at Rothamsted, where the mean rainfall (per 
day) in October is about 54 per cent. greater than 
that in April, than it is at Blue Hill, where the mean 
rainfall (per day) in February is only about 24 per 
cent. more than in June, Nevertheless, even in the 
Rothamsted data, the variability of rainfall in the 
same calendar month is so great that the mean 
values give little or no indication as to which month 
should be expected to score a record for rain or 
drought. Indeed, both records are at present held 
by December, which in 1864 gave one-sixteenth of an 
inch of rain (0°063), and in 1914 gave 8103 inches. 
R. A. FISHER. 
Rothamsted Experimental Station, 
Harpenden, Herts. 
