HABITS A S l J FEKDINC. MECHASIS>I OF RIAUSTOHIUS 4RENARIUS
363
The habits and feeding mechanism of the Amphipod Haustorius arenarius
Slabber. By RALPHDENNELL,
B.Sc., Research Assistant and Assistant
Demonstrator, Department of Zoology, University of Loeds. (Communicated by Prof. IV. GA~LSTASG,
M.A., D.Sc., F.L.S.)
(With 13 Text-figures)
[Kcntl 24 November 19391
INTRODUCTION
The sand-burrowing amphipod Haustorius arenarius aftor special search
was found in considerable numbers in Robin Hood's Bay, Yorkshire, and
a study of its structure and habits was undertaken a t the suggestion of
l+ofessor W. Garstang. Professor H. G. Cannon, after seeing one or two
dead speximens, suggested that the +ma1 probably fed by some filtering
mechanism, and accordingly the feeding-habits have been examined in soma
detail. My thanks are due to Professor 11. G. Cannon for his kindness and
;mistance.
Haustmius has been described and figured by Stirs (1895), and by Chovreux
and Page (1925). Stebbing (1906) quotes the following synonyms for the
genus Haustorim :--Lepiductylus Say (1818), Pterygocera Latrcille (1829),
Be& Spence Bate (1851), and Sulcntor Spenco Bate (1854). For tho
spccios Haustorius aremrius he gives the synonym Oniscus arenarius Slabber
(1760).
spencc Bate and Westwood doscribe the animal in their ' History of the
Hritish Sossile-eyed Crustacen,' but it is not clear whethor or not they are
confounding it with S'u,lccLtor ( =[Jrothoe) noroegicn. Stws suggests that most
of tllc?chilractcristics of the animal may I)o csplibind as adaptive modifications
in relation to tliv n d c of life. This view is amply eorroborthd by my own
observations.
Hunt (lY't5) has discussed the feeding of several amphipods, and divides
them according t o their habits into suspension-feeders and detritus-feeders.
H(: points out that there is no doubt a good deal of overlapping betweon
tllese two divisions owing to the floor of the sea being in a liibilc condition,
cletritus boing stirred up by the wator m d suapcndcd organisms dying and
becoming deposikd on tho bottom. The feeding of Arnpeliscu is briefly
&scribt+. Apart from this paper I know of no other description of feeding
in the hmphipoda.
A description of Haustorius from the point of view of its habits and feeding
will L c given later. The ttccounts given by Sars and by Chevrcux and Page
25*
are exccllont from the titxonomic Standpoint, hut tire insuffir:ient for the
purpose of discussing movements and feeding. Hizustoriirn feeds on finc
particles of detritus, which itre collected by ii c:ombintxl niaxilltLry ' suctionputnp ' and filtcr-plato. Stebbing (1906) N i L Y s of tho :inirnd thiit i t ' barrows
with great dexterity in wet s:tnd. Will live 1 ~ i gin suitd)lo confinemont. '
These romiirks nre confirrutd by my observations.
HABITS
(1)
&CURRENCE.
Hauskxizis arcmrius occurs all round the coasts of the British Tslea, the coast.s
of France, tho Katttrgat, Holland, the east coiist of North America, zind the
west coast of Norway, burrowing in the sand of the lmach (Chovreiix and
Fage, 1925).
At Robin Hood's Bay, on thc east coast of Yorkshire, eonsidorublc numbers
]lave h e n found, and the occurrence of the aniinnl here is intcresting in thiit,
although i i sand-inhabiting animal, there is vory littlc s:intl on this ptirt of tile
coast. Thc distribution of the animtal in this bny is so intimibtely bound 1111
wit,h tho physical features of the bay 3s to merit description.
T t will first be nexessitry t.0 o u t h e thc topography of the bay, whic:h is d m n t
thrct! niilns wide. At low tide the sweep of tho bay is brokctn 1)y n, serios of
rcc+, running out to soti at right-angles to tho shore (text-fig. 1). These
consist of fossilifcrous shales of lower Liiissic agc, dipping gcntly to the nortli
rirltl hrokori into low wtirs on the southern side of ccich reef. Uy fiir tllc
greater part of their surface in covcred by littoral algae or cnctrilstatl wit.11
barnacles, but in one place, about the middlu of tho bay, there are doposits
of fine clean sand between thc scars, and it is in this sand that Hautmius
is to be found. Nowhero does this sand exccetl a dopth of more than two
to three feet. (These san& tire continutilly shifting ; this observiitiori wits
made: in the summor of 1928.)
Apart from the sand lying between tho scars, thcrc iwo tracts of s;bntl
lmrdcring a stream, Mill Beck, which flows into tho bay. Zlazistorius i n also
fo1lnd in these H L L I ~ ~ R The
.
aninicil is novor found in sund which stiolvs i~11y
tt:lidcncy towtirds drying up or which contains much ( I t h i s (Table T).
By digging in vnrinus directions in tho siind it was found powi1)lo to obt:iin
some chmatc of the frequency of occurrence of the rinimal. Tho trc!ncIles
were 10 feet long and 1 foot wiclo and dcop. The sand from the trcncI1c.s
was carefully examined by passing through the I l i ~ i d ~ If
.
this is tloT1c
t,horoiip;hlyit is probablo that very few of thc I q e r animals o m p a s s d over.
The organic matter present a s detritus i n the s m d viwiw in qutintity from
plncu to pliice, and is in n finely divided condition. When obst:rved ilz sniilll
sitnd samples under the binoculur microwopt! it appotirs iis flocculont pwtioles,
very much sinaller thiin t,hc santl-grtLins. It is prol)at)ly mainly clterived fronl
t}ic attrition of the litt,oriil algae.
F E EDIXG .ME:CllASISM O F 1IAI:S:TORI U S ABEN A K I U S
3ti.3
Sand-samples wwc! tiL1mi from various positions a t the bases of, and
botwccn, the SCiLI'S, ; L I tlioir
~
o r p i i o content determined by first drying
'~'ExT-FIG.
The Scar systexii ill livbin Hood's Bay.
1.
L. =Leeds Univereity Biological Laboratory.
thoroughly and tlien igniting a t a heat just below rednesa to remove t,he
organic matter. The loss in weight aftor ignition gives an indication of the
organic content of the sand.
366
MR. XALPH DENNELL ON THE HABITS AND
TAHLET
II
/
So.
'
I
I
I
Position of 'lhiiclr.
,
~
I
1:
'I
I
At base of scar and parallel with L.W.
I
.....
,
A t haso of scar and almost parallel with
__
............
At base of ticar arid parallel with L.U'
3.
No. of
Animals.
i
-
~-
.
S a i d very wet.
4
__
__
--
.
'
Much blark dahris i n
sniI(1.
. . . . . . . . . . . . .
Much weed-deliris,
_.
ashos, coal, otc:., iir
send.
,
Same scar as above : siiniliir position,
but 10 yards away.
! 3.
--
___I
__
Ditto.
3.
~
I
Kerrrarks.
Ditto.
A t baso of scar and parallol with L.W
-. .........
lJarallel with L.W., h u t 8 feeb from base
of above Hoar soaward.
A t bum of soar and pnrollel with L.W.
.......
__
Ditto.
-:;1
,-.-
!, -..j
-
,
,
Wrt. fine sand. SfJllm
bluck dobris.
.
Parallel with L.W., but 20 feet froin haw
of nbovo scar soaward.
7.
Cluun and wet.
2
Ditto.
1
__-- - -
- .._ _ _- -
I
At baao of scar and pnrallol with L.W.
8.
1
!-i-
Ditto.
i
I
I
!
'3a. Parallel with L.W., but 10 foot from base
of above scar seaward.
I___
! 11.
,
1
I
At base of scar and parallel with L.W.
10.
12.
~
-
Ditto.
1 Ilitto.
-
1
I
Dry sand : littlo
dohris.
j
Ditto.
I
_
'I'ronchen in sand with varying amounts of debris.
For explanation, see base of Table 11.
.
I
361
E’EEUIXL: MECHASISM O Y IIAUSTOHIUS ARENARIUS
TABLETI.
KO.of
Organic content
Aniinals. ‘ofsand (per cent.).
,
l’osition of Trctich.
KO.
I
l’trrttllel to scar and near to its l i a ~............. I
72
- -_______1 a. l’ttrallcl to tronuh 1, but 4 feet nearer L.W... . . . . .
4!)
1.08
-__
i
______ I
2.
3.
-.j
I
1.
.
’
-
-- -
1’ttraIIcI to base o f scar antl near it
_--
l’arallel to base o f scar and near it
.............
3 u. I’arallcl to trexich 3, but 6 foct iicarer L.M’.
... -.
-
.............
____
-.
34
.$Hi
1.10
1.02
~
......
-
.(J4
Tlloac: trenches were dug in fino clean band, and wore oach 10 feet long arid 1 h o t wido
and dmp. Tho nurrihrrs in tho left-hand column rofor to tho distance of the
trcnchos from tlie low-water lev01 (e.g. 1 donotcs the first scar landwards from lowwater level, 2 tllc scco~ld,and so on.
Tahlc 11. shows the iiuniber of animals antl the organic content of the sand
foulld in various positions. Since the animals feed by filtering smiill foodp;irt,icles from tlic surrounding water, it is probable that variations in the
organic content of the sand will influence their local distribution. As receding
W~LV(:S pass over tho witis ib backwash current will bc produced, and possibly
this current may deposit detritus in the deep sand at the bases of the scitrs.
A4~iotherfactor influencing loud distribution of the aiiimals is the watercolltellt of the sand. It is soen from Table I that the animals are more
liiimeroiis at the bases of the scars than hetween them. The spaces between
t,hc scars resemblc long shallow troughs (text-fig. 2), and thc deepest parts
of these troughs arc situated at tho hases of tho vertical fttces of the scars.
Towards the sea tho t,roughs become shallower. Tho troughs are filled with
salld, and water left behind as the tide recedes will tend to colloct in t,he
& e P r parts a t tho bases of tho scars. Since the burrowing power of the
animal is due to the powerful current expolld by the ploopods, therefore
rendering t,he iLniin;il unable tr, burrow in dry s a d , the high frequency of
Occllrrence of the animal at the bases of the scars may be duo t o burrowing
beilig easiest t,here. Also, sinco feeding is dependent on a current of wator
producd by the maxills, it will be most easily carried on there. There
is thus it diminution in numbers of iinimitls on each scar from the base towards
the sea (text-fig. 2 , 11’).
Considering the war system as a whole, the animals are most numerous
on the scar nearest low-water mark, and arc progressively fewer on succeeding
scars towibrds high-water mark. This is due to the tondency of the animals
to congregate in the wettest sands, those nearest low-water mark, and thus
368
MH. ltALPH DEXXELL O N TllE HADITS A X D
they show a diminution i n nunibcrs on tho whole beax:li, passing l;L~ldw;trils
(text-fig. 2, Ill). There is thus a primary diminution, W, up the beach ibS
a whole, and a swonditry diminubion, lF, in t,hn rov(!rsc: clirrtction on c!it(:lL
i n ~ l i ~ i 1 ~ 1w1i ri ~. 1both being wuountcd for by the witkcr-colltent of the s m d .
The possibility that the greater protection gi v m to tJw ;miniills t i t the biL?ic?%
of thc scars may also influence tticir abundance t,hcrs is not to be ne@;ltx:td.
Huustorius app:irently may Sorsako thc! sirnrl. t i t periods of high wiztcr to
tiwini freely. Scott (1922) found indivicluals in the stomiwlls of yoirng l)Iii,i(:tt
(27-78 mm. in Icngth), and, since tho young plaice fccd idmost entirely on tho
pelagic and semi-plttgic invcrtebrat,c animals living i t i thc! w:Ltcr through
which they move, it would appear t1i;i.t Uawstorius docs leave the sand rltiring
high-wahr periods. The same author also states th:it in thc Morccitmhc l31i.y
area it tow-netting taken a Sew y m t s out from the shore, and in not more than
two Ssst of water, usually contairiR euch truly pelagic forms 8 s Calanus and
Tcmura, along with tlic bottom-living (jumacca and AnqdiiIwda (cf. ‘ Tropisiiis,’
11. 374).
‘l kXl ~-FKO.2.
&.-----------.--._
..______
.._.
- _ _ __ ..--.-..
D’
~
*
b2_ _ _ _ _ _ _ _ _ _._ - - - - - . - - - .
-
Sand
Scar
50 W C 1
The mars in tronsveree Bection.
The Hand on the scars contains u few small Nereid and Sabellid worms
and the Amphipod Sulcator (=Urolhoe) norueg,ica. With the exception of thaw
and occasionally a few sand-ools (Ammodytes b n c e o b t w ) , no other animals
have been found in the oand. Ko large miteaes of dutritus are, to be swn
in the nand, and from an examination of the sand it would seem that feeding
must take place on smdl ptwticles ritthur than on largo IntLYseB. The ibction
of the sand, set in motion under the influence of wave action, will result
in the comminution of all organic msteriitl down to a uniform fine grade
(Bruce, 1928).
(2) MOVEMENTS.
It will now be necessary to give a description of the animal from the
standpoint of its habits.
The body is broadly arched, the littertbl compression typical of so many
of tho Amphipode k i n g :tlmost completely lost. The coxal plates aro large,
and these, together with the broadly arched body, give tho animal a boatshaped appearance when swimming upside down. The joints of tho last
three pairs of pereiopods are broadly dilated into flat laminae, and these overlapping form with the coxal plates a wall on each side of the body (toxt-fig. 3).
E'EEDISO MECIIAKISM O F HAI;STOHLUS AKEKARLUS
369
Tht:re is, then, :L tiintiol-shaped s p ~ : c1)ctwet:tl those walls and the body above,
and this is of tho greatest importance in the rapid burrowing for which the
creature is rcrnitrktt1)le. The plrmtifril sct:tl iirmature of the three pairs of
iweioyods interlocking and owrlapping enhance the efficiency of this wall.
The first ;r.nd scieo~idpairs of' pwctiopods ~iossessposteriorly on the carpal
joint it rounded limwllar c:spa.nsion, atid thc: outer part of the propodial joint
is obtusely rounded, thus forming i i shovol-likc structure.
The pedunculatr joints 0 1 the antmilit: itre obtusely rounded it1 a sirnilar
manner t o the propodinl joints of the first two pairs of perc:iopods, and arc
in addition provided with il dense fringe of marginal setae, copiously feathered.
The pedunculor joints of the antennules are rounded in a similar manner,
but to a far less extent. The plcopods arc cstrcmcly powerful, their bases
'I'EXT-FIG.
3.
possessing on their inner margins :I pair of coupling-hooks, which lock u u h
p i r of plcapods together. The sctae of thc pleopods are rather stout prosimally, and not plumosc, whili: distally they aro more slender and densely
plumose. Thr: setae of each ramus of a ploopod interlock and form a doiise
mat, so that in cffcct ewh pleopod presents a flat, water-resisting surface
to the rear.
The uropods are very stout, and are directed upwards and backwards,
and provided with strong spine-like setac?.
Not only is t,he lateral compression typical of the Amphipod lost i n
Haustorius, but the flexion of the body is very greatly reduced. This flexion
probably prevents an Amphipod such i t s Gnm,marus locustu from swimming
in a regular manner, as it curved body travelling through a dense medium
such as water will tend to follow it curved path. Hau&wius is beautifully
construct.ed for swimming and burrowing.
The eyes are small and white in young specimens, but in older animals
they may become almost invisible.
370
MH. RALPH DENNELL
OX THK HABITS A N D
All the spwimenn collwtcd ;It Robin Hood's Bay wero femalos. Sars (1895)
figures only the female, and Chovreux imd Fagc (1925) state that tho rnele
is unknown.
(iL) S?uirnming.
I/uuslori?is swims on its back (test-fig. 3) in ti very gribceful mallner, parformirig c:volutions quite impowiblu to siioh an A4nipliipod ibs C:atri,narus
l o c i ~ s h . Whcroas Gamtiiarus locusta dcpends on the rapidity of its movcmcmtfi
fLr support in the water, Hausturius is i~blcto swim yuito slowly-indeotl,
it, can hover ibhnost in one placo for a considerable time. Thin power of
swimming with ease, though not rqlidity, is conferred by the broadly arched
body, tho vtsticail plates formed by the coxal plates and expanded joints
of t,hc last three pairs of poroiopods, togethcr with the setae borne by t h
;tiittrirriulcs and antennae. Thus tho ibninld prosonts a considerihlo surface-
TEXT-FIG.
4.
L _ .
--
4
.__5
t
-
6
SLUW iilg ditqprnuiaticdly tho inovcinciits of thw wtao with respect t o tho
Y K I I I U L ~of
a plcopod during its beat,. (The direction of mobion of tho I~RIIIUS ici ititiicntccl
by p thin arrow below i t ; the currelit produced is mpremuted acco~dingLo its
stroiigth by tho Icngth of the houvy arrow; a d buckwash currents UPO shown
by tho thin curved arrows.)
area to the watcr when swimming, and this IiLrge exposed iLrea, alt]1ougll
it iIlcreascs tho resistance of the m i m d to forward miotion, arid so m&:s
swiirimirig slower, also increases its resiatimcc t o dOWnwibrd ftdling through
the water. I n e h t , thc ttntcnnulcs, allt~nnae,COXibl plates, anti perciopods
serve as organs of flotation.
It hils already b w ~ iremarked that thc shape of the body when swimming
is i)erf&tly boat-like. Tho powtrfril swimming current is produced by the
rhythmical boating of the thrcu, pkri of pleopods, arid LL few rernmks on the
current production by the pleopodR will not be out of place.
I n text-fig. 4 is shown tt ramus of a pleopod in six Liiflerent positions. The
currunts produced are represented by arrows on the limb, and the &&ion
FEEDISC: MECHASISM OF HAUSTVHIUS AKIGhARIUS
37 1
in which the limb is moving is shown by the arrow below. ConsiderLig first
the third position of thc limb, it has almost reached its limit of travel ill a
backward direction. The setiw of the pleopods are so attached as to be al)lc:
to movo ovor a considerable rango with regild t o the axis of tho limb, and
it is this fact whlch is rosponsible for the efficiency of t h e pleopods as
a propelling mech;mism. The sctae in position 3 of thc pleopod hiLve bcen
forced to extend laterally from tho axis of the limb by the pressure of the
water on them, and in this position thoy offer their maximum resistance.
As the limb is moving backwards il powerful backwardly directed current
will be produced.
I n position 4 the limb has commenced its forward stroke, and thc rcsistmce
of the water, acting now on t,he anterior face of the pleopod, causes thc!
setae to stream out behind the iLxis on tho whole of the forward stroke. The
effective area of the plcopod is thus reduced, and the forwardly directed
current produced on the forward stroke of the limb is very feeble compared
with that produced on the back-stroke. Thc setiie arc seen trailing behind
the axis in positions 4,5, and 6. In position 1 the limb has begun t o travel
backwards again, and the setae aro beginning to spread out laterally again,
which they complete in positions 2 and 3.
If no provision were mado for the collapse of the pleopod setae oil the
forward stroke-that is, if both back and forward movements of the pleopod
produced currents of tho same strength-no backwardly directed swimming
current would be produced, for this current would be annulled by ail e y d l y
powerful forwardly directed current. The only result of the boating of thc
pleopods would be to produce a violent agitation of the water in the neighhourhood of the pleopods in ii series of eddy currents. The pleopods
inevitably c i ~ u ~backwardly
c
directed currcnts round the axis of' the limb
on the forward stroke, and forwardly directed currents 011 the back-stroke,
but these d d y currents are feoblc in comparison with the powerful backwardly
directed swimming Current, and do not greatly reduce the efficiency of the
pleopods as a swimming mechanism.
An arrangement which increases the efficiency of the plmpods is seen in the
coupling-hooks or spincs on the inner sides of the bases of the pleopods.
By their agency t.he bases of tho pleopods are linked together, so ensuring that
tho pairs of pleopods beat together.
So far the beating of one only of the thrcc pairs of pleopods has been
considered. A moment's observittion of the animal shows, howover, that
all the three pairs of pleopods do not perform their beat together ; they do
do so out of phase with oach other. This difference of phaso is of a particular
type-that known as metachronial rhythm, in which, in a series of appendages,
any one appendage begins its beat immediately before the appendage in front
of it (text-fig. 5). Such a rhythm is of fundamental importance in producing
a steady current instead of a wries of sharp jets of water. I n addition, the
effectiveness of each of the pairs of pleopods is increased. If all the three
Ijairs of' plco~iodsprformed tJioir h a t s togother, the current produced would
l)c lit,t.lc, if ;my, ~ ~ i o rpowerful
c
tham that prodiic:ml by thc Iitst pair idom:,
il.5 (';ich of' tho firsb two paiirs would be tritvellitig in the backwash ciirrcnt
I)roduc:cd Ly t,lie pair behind, ii condition not conducive t o efficiency in iblly
l ) i d t l l t ! SyStPIIl.
:l~tl~iblly,
illitlor t.hc infliictric:e of xuctiu:liroiiial rhythm, a pair of pleopocls
v i l l i work undisturhod by the effects of the othor pairs. Tri the case of tho
li~~
imir
t , of p l c o I d s , their back-stroke produws a powerful bnckwardly directed
currant,, h i t iis the pair of pleopods in front is Ittgging, the spitce bctwccii
thew two patirs is iricreascd. WiLt,(:r flows in to fill this SI)it(;e, and i t is this
water wliidi is set in motion hy tho bawk-stroke of t.he second pair of plcopods.
l h o 1)ostc:riorpaiir do not impede thc: outflow of tlio swimming-current produced
r ,
I
2
4
5
3
by thc: mitldle pnir, i t s they iirc at thc end of their bateli-strokc! aid are momentkirily laying itlmost flat beneath thc abdomcii. Tho smno happcns with tho
first two pairs of pltwpods.
Tho pleopods, then, working wit,h a beautiful mwharucal nicety, ~""~lucc:
;L steady powerful swimming currcrit passing l)ii&wards h n e a t h tho ;hlonieii.
The reaction of this current on t.he plttopods propels the aniinal through the
water.
Stccriiig is effected mainly by the lhst pair of pweiopods. The expanded
joints of thcst: limbs, csteiided bitc:kwar& parallel to the direction of motion,
serve ;IS A pii8ir of ' ruddt:rs ' below the plcopotls. 'l'lic position of tho
* rudders ' is of int.crest, for the water flowing pa& them is expelk?tl by the
pleopods and is travelling faster with reqect to the body than the water
passing the body morn anteriorly, when the animal is in motion. The
reaction o f t.he w d m on the steering limbs is t,hcrc!fore p u t o r than if more
ttnterior limbs performed the function of steering, and so increased control
over the swimming movements is given.
FEEDIXG m c I r A s I s M 01:
IIACSTORIUS ARENARICS
373
Possibly the large coxal plates and dilated joints of the pereiopods may serve
to prevent rolling of the body when tiit: itnimal is swimming, as any side
to side movement will be checked by thc! opposition of ;L I;iryo fliit SllrfiWe
to the water.
(b) Bwrrovhg.
The mothod of hurrowing is of interest, for i n s t o d of the sand being s c o o p ~ l
away by the digging action of the limbs, i t is expelled by the action of the
powcrful currcnt produced by the pleopods. It is true that the first two pairs
of pereiopods do serve t o R slight extent in digging, but the chiof agent is tho
cxpulsive ,action of the swimming stroam, aided to some extent by the rrropods.
As has been mentioned beforo the coxal plates and the large oxpantled
lobes of the three pairs of pereiopods form vertical walls below the body, so that
the space enclosed below the body is of tunnd-section. Furthermore, numerous
long setne borne marginally on the joints of the last three pairs of pereiopods
projoct inwards below and across this space, so that thoro is an almost cornp1t:tely enclosed tubular space below the body.
A swimming individual sinks to the surface of the sand in the tank and
i~nm~xlii~tc~ly
rights itself, so that the ventral surface is once more downwards.
Thct plcopotls cont,iniictheir action, so that a current of water passes backwards
iindt?r the l)ody, t:ntt:rs tlie ' tiinnel,' and is violently cxpellcd to the r w r .
This ciirrciit is suficiontly powerful to cause sand-grains to be c a r r i d dong
with it, so that sand is rcmoved from the region iit the front of the animal,
t~iroilghthc?' tiinnel,' and is cast oiit behind. .%]so,sinco thc: enirnal is burrowing
in nand sribmerged below wiiter, the swimming action of the pleopotls drives
t l m :uiini;al against t,he sand in front of it, so that i i tendency of tlic h0ad
to bore into the sitnd is given. The expanded second and third joints of the
:Lntc:nnsc, lying ;tt iLn iwiite angle to e i ~ d lother, with the apex pointing
forwards, may serve to deflect sand t o each side of the body.
Whm im individual is beginning t o h r r o w , the abdomen may occasionally
1~ swn to 1)c:conie flesetl downwards and then snddenly straightend. The
c.1fcc.t.of't,liis is t.o give tho stout spiny uropods a purchase on the sand, and tlicn
hy tht. str;LigIitt:ning ( J f t.hc body tho animal is pushcnl forwiirds and downwartls into tho sihnd. While these operations are being carried on, thc? first
:1nc1 sccontl pairs of pereiopods are assisting the burrowing by x:riiping tho
siintl, which, when loosenoti, will bo citrricd away by the backwardly flowing
currcmt. Tlic: cliggiiig eflect of the preiopods must bc small cornpiwed with
tht: atrtion of t.lic current in removing sand.
T t will bc: noted that the ~ n o v ~ m e nint ~swimming and burrowing itre
filridamentzdly the same, cach deyeiiclirig on the production of i i poworful
current by the pleopocts. Obviously the ' tiinnel ' effect of the coxal plates
;Lntl the third, fourth, tind fifth pairs of pereiopods contributes in litrgo
meiLstir(b to tlic: c?fficicncy of the water-ourrent in expelling sand, for, wen:
it not for this, water, indxad of strca.ming directly backwar(ls 1wrw;ttli tlie
374
MFt. RALPH DENNECL O N THE HABITS A N D
body, would stream inwards t o the pl~wpodsfrom all positions iirouncl tlie
1)iitly. Tn addition to t.his, the confining of the currcnt in IL ' tunnel ' has tho
cffect of incro:iuing the velocity of the current, i t r l t l also therefore its currying
power.
A large IIuustorim will produce a current suficicmtly powerful t o throw
the siind out behind it in the wuter to :t distance of almost an inch when
beginning t o burrow.
Jndividuitls have never been observed to burrow in sand that is not completely saturatcd with wittcr. For instance, animals p l i ~ e don the be'wh
on the surfiu:o of the sand arc incapnblc of burrowing and are completely
Iinlplesu. This is, of course, due to the fact that burrowing is dependent upon
an expelled current, and hence lack of wntw p i t s the burrowing mochmism
out of action.
( c ) Tropisms.
When swimming, the animals arc positively phototropic. A parallel beam
of light two inches wide was thrown across a tank in which ixidividuals were
swimming. When their casual movements brought them into the beam,
t,he itnimltls imnietlia,telyorientated tliemselvex RO as to face the source of light.
After swimming along tlic hmm towards the light until reaching the side of
the tank, t,hr!y would remain for some time endeavouring to swim forward,
arid then fall to tho sund and immediately begin to burrow.
Animals kept in captivity showed for some days after capturo a diurnal
rhytlim in their swimming wtivities above the surfwo of the sand. During
the day, from about 6 a.m. to 11 p.m. they remained quiescent, burrowing
beneath the sand. After I 1 p.m. until about 6 a m . individuals were to be
sem swimming freely in the wiitor above the sand. Obscrvcitionx were m a d e
during the day and night, and a t night a light was only iisetl for a short period
n t intervals to note the numbers of ariirnida swimming.
(3) BREEDING.
Tt, is interesting to noto that the male of IIairstmizis nrenarius is unknown
(Chcvreux i t r i t l Fage, 1925). Of over 120 specimens examined all were fi~niitli~s
possessing oostegitcs, or lucking them, but resembling those so provided in all
other respects.
Some of the individuals kept in the tank were found t o be carrying eggs
in t.he brood-chambers :It, t,he ctntl of February, but none of the eggs clevelopctl
into embryos. In this connection the work of Hlogvad (1934)may be ment,ioricul.
Hc fintlfi t.hat muny of thc Criimmarids arc anniiel spwios, dying in the spring,
after the embryos have become free. Some individuals breeding later in thi:
slimmer mag snrvive through the following winter i i . d so carry the specios
on from onc yoiw to the next.
IIrrtistmius also may bc tin iillnlld specics, for from the beginning of February
i~nimalsin thc titnk 1)rg;tn to die in c~onsii1wiil)lennmbers, although none
FEEDING M E C l l A S lSM O F ITAl~TSTOHI13ARENARII'S
375
of the individuals ever produced embryos. It may be that the heavy deathrate was duo to some poisonous substance which h i d accidentally entered
the tank, but if this wcre so it is not c:lt?;trwhat the substance was and how
it cmtered the tank, for care was takon t o keep the tank scrupulously clean.
Tn the spring (April) of 1!EX scv(?r:il s p i m o n s were found a t Robin Hood's
Bay, while in Jnly 1928, the animals were abundant, and again in October
of the same year. I n March 19% a great scarcity wiis apparent, only seven
specimens being taken after a lahorioiis search.
Whether rcprodnction takes place psrthenogcnetic.;Llly, or whnthr t h m ?
may bu B malt! of thi: spcxies, in unsntt,lctl.
THE FEEDING MECHANISM
(1) METHODS.
Tho filtering mechanism of Haustorills cloponds for its action on tho
intcdocking and particular arrangement with r e q m t to each other of numerous
scrim of setao, which cannot he clearly seen if ordinary mothods of dissection
are used. The method iiwd for Haustori7~swas first to embed the animal
in celloidin, followed by embedding in wax. With a sliding microtomo thc
animal was cut into thick sections or slices, about 0.3 mm. thick, which were
mounted unstained in hnlsam, thc cover-glass being supported by fragments
of broken covcr-glass. By examination of t.hcsc slices t,he arrangement o f tho
sctw and the disposition of the niouth-parts could ho clearly m<do out.
I am indebtd to Professor Cannon for showing mc his method (tlt?scribed
by Cannon, 1927).
By dissecting off the individual mouth-parts their structure was determined,
although by this mcans no r e l i d h information on their arrangement in situ
was obtained. Appendages were removed and cleard in glycerine, and
cs;kmined under the microscope.
Animals werc observed fooding by att,whing them t o tho bottom of a glass
capsullr with plasticinr, anti slipplying them with suspensions of various
substances. Suspensions of starch stained with iodine, carmine, charcoal,
iron saccharate, and mud wen: used, but by far tho most satisfactory was
a suspension of st,arch stained with ioclinc. Tha pirticles are not rojwted
by the animal, and owing to thcir deep b111ecolour are easily seen, and their
progress in t.he food-stream followed.
A very useful sonrce of information was given by leaving the animals for
a day in a suspnsion of iron saccharatc. Particles Hahere to the filter-set;ie
and mouth-parts, and, by cutting sections of such animals, the passage of foodparticles can be followed, and also the limbs and setae concerned in feoding
can I)e distinguished. For the time required for observation the animals
appear quite healthy in iron scamharate,as Cannon and Manton (1927) found
for Hvm im y k .
376
MR. RALPH DESNELL
ON THE HABITS A S T I
(2) RET~.4TIONHOF
MOUTH-PARTS.
Rcfore hsciissing the fiwlirig mcx:h;inism it will bo ncce~saryto dwcribo
the relations of thc moiith-purts to each other.
Thc mouth-opening is sitri:itcd vuntrally, and is bonnded anteriorly by tlic
large lnhrum, postcriorly by the well-dovdoped pibrugnuths, and laterally
by the molar and iiicisor processcv of the mandibles. The mouth leads almost
verticully upwards by :L short octsophnpis into a proventriciihs or crop,
TEXT-FIO.
0.
Uiagrwni based on R sagittal slice of Hawstorius, wmcwlint to t,ho riglit! of t l i ~ *
mcdinri plnno, to s l ~ o wtlir clisposit,iotiof the itiouth-parts.
providcd with a ' Triturationsapparat ' (a ring of chitinons teeth sitil:itccl
:Lnteriorly) and a gastric mill (test-fig.6). Thc lubrum projects vcntrally, and
its antcrior face is convcxly rounded, while the posterior face has the form of
: I , T ~obtuse v, with the ibpc% dircwtcd bnckwiwds. Internally are muscle-bands
:Lttaclied to thc labrrim arid to the c.hitinor~snndoskolcton. The postcrior
f w of
~ tJit! 1ill)runi is arincd with short dcnw sctiie.
FEEIIIXO MECHASISM OF HAUSTORIIJS ARENARIUS
377
The paragnaths consist of two ridges, projecting ventrally a t each side of,
and parallel to, the mid-ventral line, and with their anterior ends flattened
and closely applied to tho posterior faces of the mandibles. Enclosed betwoen
tlio ridges of tho paragnaths is ~1 shallow depression, tho ‘ food-basin ’ (Cannon
and Manton, 1927), around which are grouped tho posterior pairs of mouth-parts
(text-fig. 7, left). The tips of tho paragnaths are armed with setae similar
to those of tho labrum.
The setao of the labrum and ptra,gnaths project b,wItwards and forwiirds
rcspwtivcly over the space below tlie mouth-opening, so that the unohstructcd
ontrmcu to the nioiith lies posteriorly between tho p m q p t h s .
TEXT-FICA
7.
f i - - 1 .
(Left)Diagram to show the labrum and poragnaths of Elaz~storuisin ventral view.
Tho shadod areas roprauent tho bases of the limbs, which hsve been removod. (After
tho diagrum given by Cannon and X m t o n for Hemimysis.)
(Right) Diagram based on a transvorse slico of Haustorius, 8 little in front of tho mouth,
to show the lubrum ant1 mantliblcs, viewed from in front.
The mouth-appendages.
The mandibles (test-fig. 7 , right) are inswtecl laterally to the mouth,
and the incisor and niolnr regions project inwards between the labrum and
paragnaths. The incisor region consists of a double sharp tooth vontrttlly,
inimovnbla with respect to the remainder of tho mandiblc. Proximally to
this there is a row of blade-like spines with serrate margins. The left mandible
possesses a lacinia niobilis below the row of blade-like spines and above the
doublo tooth. Tho molar region of the mandible is rounded, and is covered
with a dcnso layer of csccndingly fine, strong setae, so that its surftwo is idmost
file-likr?.
The mandibular pnlps consist of three joints, the first being short and the
remaining two longer and almost q u a 1 in size. The penultimate joint
is armed with scattcrd, stout, spine-like setae, but tho tuft of set- a t the
tip of the ciisttd joint is composed of spines &out proximally, becoming more
CINN. JOURN.-ZOOLOOY,
VOL. X X S V l T I
26
378
MR. RALPII DENNELL ON THE HABITS A N D
slender distally, arid slightly hooked a t the tips. The distal joint of the
palp hears &out the mitltllc of its longth a compact row of short feathered
s e t m (' comh-row,' md.p.c.r., text-fig. 9).
The nwxillules (text-figs. 8 & 12, left), are situated somewhat behind the
mouth and irrn opposed in the middle line. The biting ctnditc (rnx'.b.e.)
is imiied with a row of blade-like setoc like those on the incisor portions of the
mandibles. Proximal to the biting endite is a flap-diaImd expansion, with
TEXT-FIG. 8.
Iliagram based on a traiisvorsc slice of Hairstwitla, a 1itt)le behirid tho mout.h,
to sliow tho pamjirinths arid maxillules, viewed from hcliind.
row of long spine-like sctae. Thcsc rows of sotac! project into the foodbasin space immediately behind the paragnath-tips (text-fig. 8). The
msxillulary palp is well developed, i d a t the tip is a tuft of long spine-like
set.ae. About a third of the length of the p d p from its proxirnal end is a tuft
of ;L few long featherd sctiie. TIM?
outer margin of the basal portion of the
maxillule in fringcd with closely-set short setae. The mtrxillules are closely
applied to the postrrior ftwes of the flattened pragnath-tips.
The mnxilluc (text-figs. 1 0 & 12, r#) arc attaclied some distance behind
the maxillules, leaving n space I)c:twcen these two p i r s of appendages. Eiwh
it
maxilla is composed of two 101)cs, it large delicate outer one (mx?.Z.e.)and
smaller stouter OM?. The outer thin lobes are described by Sars (18!k5)
as being semilunar in form. They we so curved as to form a concave depression
on their niorphological anterior surfwcs. The outer margins ant1 the tips
of the lobes are curved upwards, so that the two lobes, apposed t o each other,
form a scoop-like structure. Tho Inasillse do not project vertically dOWIlWibrCh,
hit, rising from thc: body just bchind the posterior limit of tho paragnathridgcs, lic almost horizontally 1)c:ricatli the food-hasin, with the tips of the
large outer lobes touching the piragnatli-tips (text-fig. 8). A space is thus
encloscd bciieath thc: food-basin, the short densely arranged sctae on thc:
outer margins of the outer lobes touching the ventral surface of the body,
outside the paragnath-ridges. I n the middle line butween the maxillae a denso
21,
m d p c r.
The last two joiiittl of tho maritlibuler pelp of H a w t o r i w , to show
tho coiiib-row of sutw, (m4l.p.c.r.)on the lest joint.
sieve of setac is formed by tlic intcrlocking of the motlerately long simplo
sctiLc fringing tlie inner margins of the inner lobes.
The maxillipcds (text-fig. 11) are inserted behind the maxillae, and lie
close beneath them almost parallel with the lower surface of the body. They
consist of a ba,sal portion, composcd of tlie proximal joints of each side ineompletely fused together, and arising from them on each side a masticatory lobe
(lobe intkrieur : Chevroux and Page). These lobcs itre armed marginally
with a row of short stiff set,;ie. The palps are composed of three joints,
the first being small, the second the largest,, and the distal joints bent a t right
angles. The proxinial joint possesses ii few long spine-like setac, while the
distal joint is plentifully iirmed with long stout serrate setae. The second
joint of the palp bears numerous parallel rows of short stiff setm arranged
380
M H . RALPH DENNELT, ON THE HABITS AND
l*c the teeth of n comb. These ' comb-rows' projcct vertically upwards
from the p l p and lie immtdiately beneath tho filter-setae (sieve) of the
maxillae. Thc ptilps of the maxillipds lie above thc masticatory lobes.
The basal joints of the maxillipeds are provided on their inner margins with
short shnrp wtae, and the tips of these endites possess similar but longer sctae
forming a denw tuft.
TEXT-FIG.
10.
Diagram batled on n tramvcrso slice of H a w t o r i u s , to show tho position
in aitu of ttw maxibe trnd msxillipods, viewed from in front.
(3) FEEDING
ON
SMALL PARTICLES.
As has been mcntioned before, the sand in which Haustorim lives contains
small flocculent particles of organic detritus. Haustmizcs fetds largely by
filtering thesc from ib stream of wiiter produccd by the mouth-parts. The
appendages concerned in producing this uurrctnt and filteriiig it are the
mil.xillae, which iwt i i s a combined suction-pump and filter-plate, while the
maxillipcds remove food-partkles from the maxillae and pass them forward
t o the mouth.
Moeements of the maxilluo
The movements of tho m;Lxilliw art' almost purcly rotatory, with the asis
of rotation lying almost on the outer margin of the large outer lobc. Thus
TEXT-FIG.
11.
Thc inaxillipeds of Haustoriuu, vicwed from in front.
011 rotation of the limb the inner margins of bhe lobos describe a wide swecp
outwards and downwards, while the outer margins of the lobes are shifted
slightly upwards and pressed against the lower surface of the body. If the
movement of the maxillae were purely rotatory, the inner margins of the
maxillary lobes would describc a11 arc of a circle, but actually they sweep
downwards at the beginning of a heat, the limb begins to rotate, and their
movement becomes dirwted outwiwds. The limbs pass through the same
positions in returning to the stnrting-point of :L stroke.
‘rhci food-basin space is contiriuelly being enlargcd irrid diminished by tlic
action of the maxillire, iind as water must flow in and out of the space as it
cliariges in volume the feeding current is set u p .
The flow of thc water entering itrid lcaving the foo(1-l)iL~i~l
is hrectetl and
controlled by the setao of the maxillae.
When the food-basin is enlarged (that is, on the outward stroke of the
Intrsillae), the voritrtil gep between tlie two mibxillae is inoroirseci, but this,
owirig to the motion of the nwxillac, tirkes place towirrds thc end of tltc stroke.
Water flows in through this ventral gap, which is c h e d by the sieve of st:t;b(:
on tlic inner margins of tlic maxillary lobes, iind food-particles ilre cleposited
on the siove-setiro. The food-basin is now full, and the return ntroke commenoc:a.
A t the beginnijlg of the strokc tho gap between the maxillae closes arid water
is unitblo to CSCUPC hcre.
The arrangement of the setac on the tips of the large outer lobes of the
mirxillae, touching the paragntLth-tips, has been de.wribcx1. Them RetiK? act
as non-return virlvcs, prcvcnting the inflow of water to the, food-basin spw9
on the outward stroke of the maxillae, but permitting its outflow on the return
stroke. Water leaves the food-basin space anteriorly through these sotac,
and food-particles which have cscaped the morginal setao between the maxillae
iiro rotaincd hcre.
The maxillae function then as a combincd suction-pump and filter-plate,
the filtering regions of the maxillae being the marginal setac between the inner
lobes and the setire a t the tips of the outer lobes.
The movement of the milxillipeh is latero-medid, the parallel comb-row
of setm on the second joint of the palp working over the sievo of sctiie on tho
inner margins of the maxilliw, and food-particles are transferred from the
Iniixillae to these comb-rows. Thc diHtal joint of tho maxillipcd-pall), tvmt
a t right-tingles, is capable of being bent still further inwards so that its serrate
setae pass over the comb-rows of the second joint and carry food-particles
from them. The distal joint of the palp is then bent forward agirin and the
particles removed by the maxillules, probably by the sotae on the flap-shaped
expttnsion proximal to the biting endite. Food-particles retained by the setae
a t the tips of the maxillary lobes are removed directly by the maxillules.
The maxillulcs work with a latero-medial movement, their upproximation
resulting in thr: tips of the sctac of the biting endite being pushed forwards
through the gap between the parab.nath-processes, Rince the setae do not
interlock to any g e a t extent. Food-particles on these setae are removed
by the setae on the incisor regions of the mandibles.
It will be remembered that t,he left mandible only possesses a lacinia mobilis.
I n this connection Cannon and Mirnton (1027) state that the asymmetry of the
mantlibles is of functional assistance in passing food to the mouth, and a description of how this takes place in Nem.imysis is given (p. 235). It appears
that food is c:iwried up the omophagiiti by 11 sucking action of the crop, for,
011 feeding a young aiiimiLl 0x1 starch-particles stained with iodine, after sorue
FEEDING MECHANISN OF HAUSTORITTS A R E N A R I I ~ S
383
time thc crop may bo seon through thc semi-transparent chitin to bccomc
suddenly full. This has been noted by Hart (1930) in the amphipod Corophium
volutator (Pallas). I n Haustorius, as well as in Corophium, the ocsophagus
is short, SO that it may reasonably be supposed that poristnlvis of the oesopliagus
is not the cause of movcmcnt of the food into the crop. Since in Criistaccans,
whatevor their fceding-methods, the mouth-appendages do not lie inside the
mouth but surround it, a sucking iiotion up the ocsophagus would carry foodpmticlcs towards tho mouth from adjacent rogions, md loss of striiy foodpiwtic:los woiiltl he decrcasod
(4)FEEDING
Oh’
LARGE FOOD-“ASSES.
Scveral points in t h r structure of the mimid suggcstcd that feeding on large
food-masses might take place to some extent. The second gniLthopods arc
cheliLte, but the c:hcl;rc :Lrc very small and hidden in the setae of the limb.
Tho serrate wtav on tho pdps of thc maxillipeds suggest that the maxillipccls
might be used for holding mnsscs of food. The mandiblcs, maxillules, anti
maxillipeds rescmblc t o some extent those of a raptorial Amphipod suuli :IS
Gummarus locirstn. Accordingly scvcral animals were starved for some ditys,
and then pl~tccdin tubes contii.ining pieces of boiled white of ogg stained with
cochineal in w;i-wiLter. Aftcr about 21 horns the cochineal could btt s ( x m
throiigli the chitin to have r c d i t ! d the dimcmtiwy system. Xxanlination.
384
MR. RALPH DENNELL ON THE HABITS AND
however, revealed that the cochineal was lying in the hepato-pancreatic lobes,
and not in the alimentary canal.
Crissans (1904) states that ' in Gammarus which had been fed upon carmine
for a week no granules were visible in the divcrticula of the gut, though the
mid-gut contained a quantity of the substance.' At any rate, the presence
of cochincal in the gut-diverticula of Haustorius would seem t o show that the
animal had been feeding on the solid white of egg. No animal was ever actually
observed to feed on largc food-masses, but i t is probable that the food is held
by the maxillipcds and mandibular palps, and bitten into by the mandibles
and the biting cnditos of thc maxillules.
DISCUSSION
It has been shown that the Amphipod Haustorius arenarius is able t o f e d
on small food-particles suspended in the water in the sand in which i t lives.
This is carried out by a filter-mechanism formed by the series of mouthappcndages, the maxillae producing and filtering an anteriorly directed current
of water.
Cannon and Manton (1927) have shown that a n anteriorly directed feedingcurrent occurs in Hemimysis, produced by the action of the maxilla, and
it will therefore be of interest t o consider how far Haustorius and Hemimysis
correspond in their filter-feeding mechanisms, and what is the significance
of any correspondence or divergence betwecn thc two.
I n Haustorius the endopodite and exopodite of the maxilla have disappeared
completely (text-fig. 12, right), and the movement of the limb in producing
the feeding-current is such that it involvcs rotation of the limb about its own
axis, while Cannon and Manton (1927) (and private communication from
Professor Cannon) consider that in the Mysid the endopodite and exopodite
are of paramount importance in the production of the feeding-current acting
as valves ; moreover, the movement of the limb resembles that of the first
trunk-limb, but is more lateral (p. 229).
Apart also from the differencos in the form and movement of thc maxillae
of IIaustorius and Hemimysis, the two feeding-mechanisms differ in the manner
in which the food-particles are rcmoved from tho filter-setae of the maxilla,
and passed forward to the mouth. I n Haustorius the second joint of the
maxillipedal palp bears several '' comb-rows " of setae which scrape across
the maxillary filter-setae, but in Hernimysiv food is scraped off the maxilla
by the endite on the protopodite of the first thoracic limb.
Cannon and Manton have suggested that the primitive feeding-habit of the
Malacostraca was filtratory, and give a diagram of the mouth-parts of a
generalised Malacostrecan (1929, p. 176). They point out that in the higher
forms of the Syncarida, Peracarida, and Eucarida this method of feeding has
been lost, and a raptorial method of feeding substituted by the suppression
of the proximal regions of the mouth-parts forming the filter-mechanism,
arid the elaboration of the distal regions into raptorial structures. For