Anal and Oral Intake of Water by Crustacea

Rectal swallowing of water in Crustacea seems first to have been observed by Lereboullet(1850). He saw it in Limnadia, Daphnia and newly hatched crayfishes. In the latter he watched carmine particles taken into the rectum and expelled again. Lereboullet interpreted the phenomenon as one of anal respiration, without giving any reasons for this opinion. Presumably it seemed to him that the entrance and exit of water into and out of the gut of an aquatic animal could have no other function.

Weismann (1874), in his thorough and beautifully illustrated study of Leptodora, described the swallowing of water through the anus, the rectal dilator muscles acting as a suction pump and the circular muscles as a force pump. Carmine particles were taken in. He, too, stated that the function of the intake of water is respiratory. He pointed out, moreover, that water is also swallowed rhythmically through the mouth, and he considered the whole gut to be respiratory.

">Siedentop (1930) made experiments with Leptodora to test Weismann’s view. If the anal intake of water has a respiratory function, its rate might be expected to increase in water deficient in dissolved oxygen, as has since been found for the respiratory organs of various Crustacea (Fox & Johnson, 1934; Johnson, 1936; Peters, 1938; Walshe-Maetz, 1952). In Siedentop’s experiments a low oxygen content of the water doubled the rate of swallowing movements.

I have confirmed and extended Weismann’s observations on Leptodora kindti (Focke) by a study of forty-nine individuals at 20−23°C. (Fox, 1952). All animals, although examined as quickly as possible after capture, had empty stomachs, the walls of which showed continuous and vigorous antiperistalsis, with a mean rate of 14 contractions per minute. The anal intake of water by rhythmic antiperistaltic movements of the rectum was clearly seen. In some individuals it went on for long periods, with few interruptions, while in others it occurred only in occasional bursts of rhythmic rectal contractions. When more or less continuous, the swallowing movements were interrupted at intervals by a pseudo-defaecation of water, which got rid of at least some of the water that was pumped in. In a typical case the mean rate of rectal swallowing was 16 gulps per minute, and four successive pseudo-defaecations occurred after 25, 38, 27 and 34 gulps. Usually the swallowing movements stopped for a few moments after each defaecation, in other cases their rate doubled just before defaecation and slowed just afterwards.

Lereboullet (1850) had found that a carmine suspension is swallowed by the rectum of young crayfishes, and Weismann (1874) repeated this with Leptodora. I have been able to show that various other Crustacea can swallow suspensions through the anus. For example, a young individual of Lepidurus sp., 7 mm. long (hatched from New Zealand mud), was watched with a low-power microscope beneath a supported cover-slip. After defaecation it often happened in these circumstances that the faeces remained close to the anus; portions of the faeces were then frequently taken back into the intestine. Another way of showing the phenomenon is to put animals into a suspension of Chlorella. These green algal cells were seen to be taken into the hindgut of metanauplii of Triops cancriformis (Bosc). In the cladoceran Sida crystallina (O. F. Müll.) with an empty intestine the cells of Chlorella were seen to be ingested through the anus and later to be re-expelled. The rate of swallowing was, in three individuals, 43, 46 and 47 gulps per minute at 18°C. In one individual the numbers of gulps between successive expulsions of Chlorella were 11, 13, 17, 14, 17, 17 and 19, in another 45, 22, 26, 21,15 and 56. These expulsions of rectally swallowed Chlorella really correspond to normal defaecations, for in another individual, which had food in the gut and had been given no Chlorella, the numbers of rectal gulps of water betweeen successive expulsions of faeces were of the same order of magnitude, namely, 37, 25, 30 and 36. In Hemimysis lamomae (Couch), too, the cells of Chlorella, suspended in the sea water, were seen to be taken into the rectum, and, in individuals with an intestine devoid of faeces, the algal cells were quickly passed forwards as far as the front of the abdomen by the antiperistaltic contractions of the intestine wall.

I have observed rhythmic swallowing of water through the anus in the majority of small and transparent Crustacea examined, including the following, in addition to species already mentioned: Branchipus stagnate (L.), Chirocephalus diaphanus Prévost, Artemia salina (L.) in sea water, Limnadia lenticularis (L.), Leptestheria mayeti (Simon), Diaphanosoma brachyurum (Liéven), Penilia auirostris Dana, Daphnia spp., Moina brachiata (Jurine), Bythotrephes longimanus Leydig, Calanus gracilis Dana, Temara stilifera Dana, Asellus aquaticus (L.), Sphaeroma sp., Ligia mediterránea F., Corophium volutator (Pallas), the isopod Astacilla deshayesi Luc., the amphipods Phronima atlántica Guér. and Phtisica marina Slabbor, Euphausia sp., various prawns named below, and zoaea and megalopa larvae. I have not, however, been able to see the phenomenon in Evadne nordmanni Lov., E. spinifera P. E. Müll., Podon leuckarti Sars, Heterocope saliens Lilljb., Diaptomus sp., Cyclops strenuus Fischer, Argulus foliaceus (L.), or Gammarus spp. A varying number of each species was studied.

In the smallest Crustacea the rectal swallowing of water is continuous. This is so both in small adults and the young of big forms. In larger animals there are pauses between bursts of swallowing movements. The following are examples. In metanauplii of Triops and in half-grown animals of this genus, and again in Sida, the water-swallowing is continuous, interrupted only momentarily by the frequent defaecations. In Hemimysis there are bursts of from two to three up to a couple of dozen successive rectal antiperistaltic contractions, with intervening still periods of irregular length without any pumping. In big prawns the water intake seldom occurs except before and just after defaecation. We have here a series of animals of increasing size: the larvae of Triops are almost microscopic, the length of Sida is 3. mm., of half-grown Triops 7 mm., of Hemimysis 10 mm., whilst prawns are the largest Crustacea transparent enough for study. Again, within one small species of prawn, Hippolyte prideauxiana Leach, rectal swallowing of water is continuous, or nearly so, in very small animals, whether the gut is full or empty of food. In bigger members of this species rectal swallowing is rare with an empty gut, and with food in the gut it only occurs at the time of defaecation. The critical length of animal for continuous rectal water intake at 22°C. was found to be 5 mm. At 4 mm. and below the swallowing is continuous, at 6 mm. occasional, and above this length, up to 10 mm. (the length of the biggest individuals of this species transparent enough for study), it is rare except at the time of defaecation. Eighteen individuals were studied.

The rate of rectal swallowing of water is moderately constant in the case of continuous intake by small individuals, but it is very variable in bigger animals and may change in a single individual from 20 to 120 gulps per minute.

From what has been written above, it is obvious that there is some relation between rectal swallowing of water and defaecation. This becomes clearer from a further study of prawns. Periclimenes scripta (Risso), a Mediterranean species, is particularly suitable owing to its great transparency. The length of individuals studied was about 2·0-2·5 cm. At most times the rectal wall is seen to be quite still. Then rectal antiperistalsis with resulting water intake may begin, at first slowly, then more and more rapidly, with finally about one contraction per second, until, after some 20-40 pulsations, defaecation occurs. This series of events may be repeated 4 times in 10 min. at 22°C. Each time, after defaecation has occurred, up to five further quick rectal inward pumping movements follow, as if to replace with water the volume of the extruded faeces, and then all is still.

In this species each gulp of water pumped inwards by the rectal swallowing movement can be seen to be passed right up to the front end of the abdomen by a corresponding antiperistaltic contraction of the intestinal wall. When defaecation occurs, without there being any arrest of the rectal and intestinal antiperistalsis, a cylinder of gelatinous faeces which was lying at the front end of the intestine quickly descends the whole length of the abdomen and is extruded. What causes the descent of faeces cannot be seen owing to the opacity of the thorax; it could be brought about by a contraction of the gut wall anterior to the piece of faeces. The descent of faeces is rapid ; in one case it took 8 sec. for a piece 1 mm. in length to move 12 mm. from the anterior end of the intestine to the anus. The cylinder of faeces passes down the intestine with some water just before it and just behind it, thus elongating the swelling in the otherwise narrow intestinal tube.

It is clear that the process of pumping water in through the anus is really a natural enema. Each time that rectal inward pumping of water starts it can be prophesied with certainty that defaecation will take place. The events described only occur if there is food in the gut ; when the gut is empty there is, in these prawns, almost always a complete absence of rectal or intestinal contractions. If such fasting prawns are given food, faeces may be expelled a quarter of an hour afterwards, always with rectal swallowing of water.

In the prawns Leander adspersus (Rathke) and L. squilla (L.), just as in Periclimenes, intestine and rectum are usually still, until defaecation is preceded and accompanied by the rectal intake of water. In Leander each defaecation is followed by a dozen or so swallowing movements, which are quicker than before the event. Sometimes, but not usually, single rectal swallowing movements occur when the intestine is empty of faeces ; this may even be repeated rhythmically at long intervals of about 10 sec., each rectal movement being followed at once by a single intestinal antiperistalsis. Occasionally, but rarely, intestinal antiperistalsis was seen without any rectal swallowing. These observations were made at 22°C. on half-grown animals, 2·0−4·5 cm. long bigger ones being too opaque for observation.

The little prawn Hippolyte prideauxiana is somewhat different from Periclimenes as regards defaecation, in that the cylinder of faeces, instead of descending the intestine quickly, moves slowly along it and is more than half-way down before the pumping in of water begins. The intestine is widest just in front of the rectum, becoming progressively narrower forwards ; at the front of the abdomen the faecal strand fills the intestine, at the hind end it does not nearly fill the lumen and there is plenty of room for water to flow forwards past the faeces. Sometimes, when the strand of faeces has passed partly out of the anus, the rectal pumping inward of water stops ; then the exit of the faeces stops too, until rectal swallowing has begun again, when defaecation recommences. This may occur several times before the faecal strand has passed the anus. In watching this it looks as if the water which is pumped into the intestine makes the faeces come out. With Hippolyte varians Leach, too, this is the impression gained. In Hippolyte there is, once again, a quick short burst of vigorous rectal and intestinal antiperistalsis immediately after defaecation. In this genus intestinal antiperistalsis usually accompanies rectal swallowing, but the contractions in the rectum and intestine are not always in step with one another as in Periclimenes and Leander, although the intestinal movements are invariably accelerated when rectal gulps become more rapid (they may reach a speed of 2 per second). Rectal and intestinal antiperistalsis can, however, each occur without the other, and there is often antiperistalsis at the front end of the intestine with none behind. The wave-length of the intestinal movement decreases progressively as the wave passes forwards.

In the fresh-water prawn Atyaephyra desmaresti (Millet), of 2−3 cm. in length, rectal and intestinal antiperistaltic waves were often seen with an empty intestine. Six animals were studied at intervals during 4 successive days. On 42 occasions rectal swallowing of water was seen, while on 19 occasions there was none. Out of the 42 occasions the intestine was empty on 18. This was unlike the behaviour of the marine prawns dealt with above, in which rectal pumping rarely occurred without defaecation. As usual, however, the pumping in Atyaephyra became quicker before defaecation and remained quick just after it, when the gulps of water were bigger.

Yet the anal intake of water with an empty intestine seems not to be a speciality of fresh-water prawns, for in the Italian fresh-water race of Palaemonetes varians (Leach), seven out of eight animals studied showed no rectal antiperistalsis without defaecation.^* The English race of P. varians, observed both in sea water and brackish water, again showed no anal intake of water without defaecation ; at nearly all times rectum and intestine were still.

In prawns, as already described, after the rectal swallowing of water has started, it becomes more and more rapid until defaecation occurs; it continues for a few moments after this and then it stops. In adult Artemia salina (in sea water) rectal swallowing occurs only just before defaecation and there is none after the event, but in younger Artemia there is continuous swallowing which only stops for a few moments after defaecation, while in the metanauplii the swallowing does not stop but it is more rapid and vigorous just before defaecation. In metanauplii of Leptestheria mayeti the otherwise continuous anal water intake, with a rate of about 40 gulps per minute at 18°C., pauses after each defaecation. Likewise in metanauplii of Limnadia lenticularis the swallowing movements, with a rate of 75 to the minute at 23°C., were continuous except for a pause of about 20 sec. after each defaecation, which took place every 1−2 min. It is curious that the rectal swallowing of water should continue for a short time after defaecation in prawns, whereas, on the contrary, it stops after defaecation in phyllopods.

Why is it necessary for Crustacea to have this peculiar mechanism of rectal pumping in of water to bring about defaecation? Why must the circular muscles of the gut wall be stretched hydrostatically for the animal to defaecate? The reason may be, in part at least, that Crustacea have an exoskeleton and therefore no muscular body-wall. We ourselves assist defaecation by the contraction of abdominal body-wall muscles, and so do birds, but a crustacean has no such muscles.

The events in prawns make it seem certain that in these animals the rhythmic intake of water through the anus is of the nature of a natural enema. The same strong impression comes from observing other Crustacea. Defaecation was watched, for example, in a zoaea larva from the Mediterranean plankton. It is unusual to see defaecation in plankton larvae because their gut is nearly always empty by the time that they are examined in the laboratory. As the faecal strand slowly crept down the intestine, the rectal swallowing of water and intestinal antiperistalsis became quicker and more continuous. When defaecation took place there was a violent burst of quick rectal and intestinal antiperistalsis, and after defaecation this continued for a moment and then stopped. In metanauplii of the conchostracan Eoleptestheria ticinensis (Criv.) rectal swallowing stopped after each defaecation, then after a pause it started feebly, becoming increasingly vigorous, with bigger gulps, until the next defaecation. In metanauplii of Limnadia lenticularis no actual gulps could be seen in the rectal antiperistalsis, but water is nevertheless swallowed and it visibly distends the posterior portion of the midgut more and more until defaecation occurs.

The spells of anal water intake in small Crustacea become progressively more frequent as the size of animal decreases; the pauses between bursts of swallowing become shorter, although the rate of the rhythmic rectal contractions does not increase. In the smallest animals there are no pauses and the rhythmic swallowing of water is continuous. This shortening and final disappearance of pauses goes together with more and more frequent defaecations as the size of animals diminishes. Doubtless the greater frequency of defaecation in small crustaceans corresponds to their higher rate of metabolism (Hotovy, 1938; Weymouth, Crismon, Hall, Balding & Field, 1944; Zeuthen, 1947) and consequent necessity of eating more. In the small animals, with continuous rectal swallowing of water and frequent defaecation, there is no reason to think that the function of the water intake as an enema is not the same as in prawns.

Water may, however, also be taken in through the anus when there is no food in the gut, or at least in the hindgut. This, as we have seen, occurs regularly in small animals and occasionally in large ones. Since the water acts as enema, why should it be pumped inwards in the absence of faeces? It may be that in nature there are nearly always faeces in the intestine. Animals are examined in a laboratory some time after capture, when no doubt food is no longer available. The rectal inward pumping of water, evolved to deal with an intestine normally containing faeces, may well continue automatically with an abnormally empty gut. But the continuous anal drinking by small Crustacea, when the gut is empty, may indicate that the swallowed water has a second independent function, additional to that of an enema. If so, what could this function be?

Lereboullet and Weismann thought that the absorption of water through the anus has a respiratory function.^*Siedentop (1930) considered that he had proved the truth of this view by showing that in Leptodora paucity of dissolved oxygen in the outside water increases the rate of rectal inward pumping. But I cannot agree with the opinion that the anal intake of water is respiratory. My reasons are the following.

First, I have tried to confirm Siedentop’s experiments, but have failed to do so. Using, as he did, Leptodora kindti, I could get no increased rate of rectal antiperistalsis in any of six animals by putting them into water through which nitrogen had been bubbled to give various deficits in dissolved oxygen. Nor, in four other animals, could I increase the rate with carbon dioxide. I had similar negative results on the rate or volume of the rectal swallowing of water by Bythotrephes longimanus, using nitrogen (eight animals) and carbon dioxide (seven animals). Moreover, Daphnia obtusa Kurz did not increase its rate or volume of anal swallowing when enclosed in a drop of water with no air surface, under which conditions the oxygen was progressively diminished and carbon dioxide increased by the metabolism of the animals themselves.

Secondly, the quantity of water taken in through the anus has the appearance of being insignificant compared with the amount flowing continuously past the gills in animals where these exist, as in Decapoda, Isopoda and perhaps Branchiopoda. This was particularly striking in small and therefore translucent isopods, where the rectal inward pumping of water and the gill movements could be observed simultaneously. In a typical case a young individual of Sphaeroma sp., 2·5 mm. long, swallowed water through the anus continuously and slowly over a period of half an hour with one gulp about every 35 sec. while a continuous rapid stream of water passed over the vibrating pleopods. The size of the rectal gulp of water was very small indeed compared with the area of the pleopods.

Thus the anal intake of water does not seem to have a special respiratory function, although, of course, the oxygen dissolved in the relatively small quantity of water swallowed must be used in respiration by gut epithelial cells. It seems probable, however, that the main oxygen supply of these cells must be from the respiratory organs through the blood stream.

Two other possible functions of anal water intake have suggested themselves. One of these is salt absorption. Fresh-water animals necessarily have body fluids richer in dissolved salts than the water in which they live. They are therefore obliged to collect ions from the water and to continue to do so, since they cannot avoid losing some ions. If salt collection were the function of the anal intake of water, the swallowing should be more developed in fresh-water than in marine animals. But this is not so. In the marine cladoceran Penilia avirostris there is anal water intake just as there is in the nearly related fresh-water species Diaphanosoma brachyurum and Sida crystallina. The marine copepods Calanus and Temor a swallow water through the anus, but the fresh-water genera Cyclops, Diaptomus and Heterocope were not observed to do so. Putting the brackish-water isopod Corophium volutator into sea water and into fresh water does not alter its rate or periods of rectal pumping. Further evidence against an osmoregulatory function was supplied by the brine shrimp, Artemia salina. Three-day metanauplii were studied, derived from parents reared in sea water of normal, of triple and of half concentrations. In each of the three waters, ten larvae were studied. For each larva three counts were made of the time (to 0·1 sec.) for ten rectal swallowing movements at 17°C. The mean times in seconds with their standard errors were: in sea water 22·6±0·3, in brine 23·2 ±0·7, in brackish water 22·2±0·9. Clearly there is no significant difference in the rates, nor was there any visible difference in the volumes swallowed. There is thus no evidence for an osmoregulatory function for the anal water intake. Nevertheless, in animals with a relatively impermeable exoskeleton, salt entry into the body must take place at least partly through the gut wall.

Another possible function for the continuous rectal pumping in of water by small Crustacea, even with an empty gut, might be to maintain the turgor of the body, thus keeping the not very rigid exoskeleton stretched. No doubt turgor is largely maintained osmotically; in Daphnia magna Straus, for instance, the blood has an osmotic pressure well above that of the outside water (Fritsche, 1917-20). But if turgor is kept up not only osmotically but also hydrostatically by the rectal pumping inwards of water, then the rate of pumping or the volume taken in should vary at different stages of each instar, since the exoskeleton is softest just after the moult. In Daphnia an instar lasts, at room temperatures, from 2 to 3 days, and the stage in the instar at any given moment can be told from the stage of development of parthenogenetic embryos in the brood pouch, since eggs are laid in the pouch just after the moult and the young swim away just before the next moult. A thorough study was made of variations in rectal water swallowing movements in relation to stages of the instar of D. obtusa but no correlation was found as regards the rate. The volume of water swallowed in Daphnia can be seen to increase when the rate of swallowing augments, but no relation was found between volume and instar stage. This hypothesis was therefore abandoned.

It seems, thus, that while the anal intake of water by Crustacea acts as an enema, it does not serve for respiration, osmoregulation or the maintenance of turgor. In prawns, antiperistaltic contractions of the long intestine carry the swallowed water forward to the front of the abdomen and continue to do so until defaecation occurs. In Periclimenes the intestinal antiperistalsis nearly always depends strictly on the rectal swallowing: each rectal inward pumping movement initiates an intestinal antiperistaltic wave. In Leander the same is usually true, although occasionally the intestinal movement was seen without the rectal. In Hippolyte intestinal antiperistalsis generally occurs only with rectal swallowing, although the two are not always in step ; yet the former accelerates when the latter does so. Sometimes in Hippolyte intestinal antiperistalsis was seen without rectal antiperistalsis, but it was then less vigorous than usual ; often, however, antiperistalsis occurred at the anterior end of the intestine with none at the posterior end. In decapod larvae all states of dependence and independence of intestinal and rectal antiperistalsis were seen. Again, in the marine copepods Calanus sp. and Temor a stilifera each swallowing movement of the rectum was continued forwards as an intestinal antiperistalsis right up to the thorax ; the mean rate in T. stilifera was 26 per minute at 23°C. In a few Crustacea, as Triops and Gammarus, intestinal antiperistalsis was not seen, but in Triops there are antiperistaltic contractions in the thoracic part of the alimentary canal. Thus rectal swallowing usually causes intestinal antiperistalsis.

The action of the water pumped into the gut by the rectum, both as an enema and as an initiator of intestinal antiperistalsis, seems to be due to the contraction of the circular muscles of the gut wall in response to stretching by hydrostatic pressure. Bayliss & Starling (1899) showed that distention of the colon in a rabbit results in peristalsis, and no doubt this is the mode of action of a human water enema. Similarly water pumped into the crustacean gut through the anus results in muscular contraction causing defaecation, with a reflex relaxation of the rectum, which is normally closed like a sphincter except for the gulps of swallowed water which pass periodically inwards.

The action of the water swallowed through the anus in initiating intestinal antiperistalsis recalls the well-known properties of molluscan heart muscle (Ranson, 1884; Biedermann, 1884; Straub, 1901; Carlson, 1906; Fredericq, 1913; Koch, 1917 ; Dubuisson, 1930, and others). If a molluscan heart is emptied of blood, it usually stops contracting, or it only contracts feebly. If it is filled again with fluid, contractions start, or increase in strength. The greater the pressure of liquid in the ventricle, the greater is the strength of contraction, and also the faster is the rate of beat. In vertebrate animals, too, the energy of muscular contraction is proportional, within limits, to the length to which muscle fibres are stretched. Starling’s ‘law of the heart’ states that increase in diastolic length of ventricular muscle fibres, distended by additional blood, increases the energy of subsequent systolic contraction. In the prawn’s intestine, likewise, hydrostatic pressure due to water pumped inwards by the rectum initiates and maintains rhythmic antiperistalsis of the intestinal wall.

After defaecation in prawns a few more rectal swallowing movements continue, as if to fill with water the space previously occupied by faeces. This may be necessary in order to stretch the gut wall just enough to maintain the tone of its muscles, without starting intestinal antiperistalsis again, which has now stopped and does not recommence until the next burst of rectal water-swallowing, preceding the next defaecation.

Weismann (1874) described the rhythmic swallowing of water through the mouth of Leptodora, its pumping inwards by the longitudinal and dilator muscles of the gullet, and its passage down the long ‘oesophagus’ to the ‘stomach’, brought about by peristalsis of the oesophageal wall.^* He interpreted the oral, just as the anal, intake of water in Leptodora as respiratory. At times, particularly with an empty stomach, he saw that water is moved forwards in the oesophagus, antiperistaltically.

The oral intake of water by L. kindti can only be seen in a side view of the animal. It consists of a rhythmic peristaltic swallowing movement in the short gullet, which runs dorsally from the mouth. In most individuals studied I found this drinking to be continuous, rapid and vigorous, the gulps of water being big. The mean rate of swallowing movements in six animals was 102 gulps per minute at 20° C. The water is passed down the long oesophagus into the stomach by a much slower peristalsis, which had, in ten animals, a mean rate of 5 movements per minute. The water can easily be seen entering the stomach because of its slightly different refractive index on arrival. Occasionally an antiperistaltic movement occurred in the oesophagus.

This continuous rhythmic oral intake of water is to be seen in the majority of Crustacea transparent enough to show it. The rhythm is much more constant than that of rectal swallowing. I have found rhythmic oral swallowing of water in the following species: Triops sp. (young, 2·5 mm. long) (78) ^†, Branchipus stagnalis (25), Limnadia lenticularis (27), Leptestheria mayeti (86), Penilia avirostris (37), Diaphanosoma brachyurum (29), Daphnia hyalina Leydig (24), Moina brachiata (43), Bythotrephes longimanus (95), Argulus foliaceus (86), Gammarus sp. (30), the mysid Siriella armata Claus (100), Periclimenes scripta (21), Hippolyte varians (33), Palaemonetes varians (21), and in decapod larvae. In Artemia salina (in sea water), Argulus foliaceus and Palaemonetes varians the rhythmic swallowing of water through the mouth was continuous in some individuals, intermittent with still pauses in others, and absent in yet others. It could not be seen in Podon or Evadne, but the gullet is not clearly visible. The rhythmic oral water-swallowing was absent in young Asellus 1·25 mm. long (adults are too opaque to examine) and in the copepods Calanus sp., Temora stiHfera, Diaptomus sp., Heterocope saUens and Cyclops strenuus, but was observed, as frequent bursts of rapid peristalsis of the gullet, in one individual of Eucalanus sp. A varying number of each species was studied.

In all the above-mentioned forms which show oral drinking—continuous, intermittent, or at least present in some of the individuals studied—anal water intake is also found, with the exception of Argulus and Gammarus. In three of the forms in which no oral swallowing of water was observed, there was anal intake, namely in Calanus, Temora, and Asellus, but in Heterocope, Podon and Evadne neither oral nor anal water intake was seen. Thus, except for the last-mentioned forms, in all the Crustacea studied water is pumped into the alimentary canal from both ends, or from either end, continuously or intermittently.

A curious detail in the rhythmic oral drinking by Daphnia and Limnadia is that each gulp passed down the gullet corresponds to a movement of the jaws. Usually water is swallowed at every second jaw movement, but if for a while the mandibles are grinding quickly, water is swallowed at each third or even fourth movement, whereas if jaws are moving slowly, water is swallowed with every movement. Yet, in spite of this regularity, swallowing is not inseparably linked to jaw movement, for sometimes the jaws stop, but the rhythmic drinking continues uninterrupted.

Greater quantities of water seem to be taken in through the mouth than through the anus : the gulps are bigger. This is the impression gained in spite of the rate of swallowing being often more rapid in anal than in oral water intake. For example, in one individual of Daphnia hyalina at 24°C. the rectal swallowing rate was 43 and that of the gullet 23 per minute.

If the anal intake of water is not respiratory, there is no reason to think that the oral intake has this function, in spite of what Weismann supposed. In filter-feeding or vortex-feeding Crustacea it might be thought that a continuous rhythmic swallowing of water through the mouth was part of the feeding mechanism. Since the limbs are more or less continuously collecting unicellular algae or detritus suspended in the water, this food must be swallowed continuously. In Limnadia lenticularis, however, the limb movements can be temporarily stopped by mechanical shock if the microscope slide on which the animal lies is dropped on the stage of the microscope, yet when this is done the swallowing movements continue undisturbed; thus limb and gullet movements are independent. Again, with decapod larvae the limbs can be held still by a cover-slip of a compressorium without disturbing the regular swallowing movements of the gullet. Moreover, the continuous rhythmic drinking is found in predators, for instance Leptodora and Bythotrephes, and in scavengers such as prawns, just as much as in plankton feeders.

Thus most Crustacea pump water into the gut both through the mouth and through the anus, or through one or other of the two apertures. We have deduced the function of the rectal swallowing of water, namely to stretch the wall of the gut; rhythmic oral drinking may well have a similar function of distending the gut wall with the result that its muscles act effectively.^* But why some copepods require neither oral nor anal drinking is unexplained.

In Cladocera (other than Leptodora) the whole midgut, between foregut (gullet) and hindgut (rectum), is often called the ‘intestine’. This is clearly different in function as well as origin from the ‘intestine’ of Decapods, which is a hindgut passage for faeces from midgut to rectum. In Cladocera the intestine, with the food confined within a peritrophic membrane, serves for digestion of food, absorption of digested products and the preparation of faeces. The walls of the cladoceran intestine can be seen to undergo antiperistaltic^* contractions which must serve to mix food and enzymes. In Daphnia, Bythotrephes and other Cladocera there is no relation between the intermittent rhythmic swallowing of water by the rectum and the intestinal antiperistalsis. In B. longimanus, for example, intestinal antiperistalsis was seen to be fairly regular with 46 movements per minute at 24°C., while the irregular rectal antiperistalsis varied between 120 movements per minute and no movements at all. In Daphnia intestinal antiperistalsis is usually continuous, although in a given individual the waves vary much in amplitude at different times. In the posterior part of the intestine the wave-length is short and the rate of contraction rapid, in the middle part the wave-length is long and the rate is less rapid, while in the anterior part there are no waves. In D. hyalina the mean number of antiperistaltic intestinal waves in six individuals was 106 per minute in the posterior region and 45 in the middle region of the intestine at 23°C.

The peritrophic membrane containing the food is, in the anterior part of the intestine of Daphnia, separated by some distance from the intestinal wall, the gut lumen here being wider than it is farther back, where the membrane touches the wall. Unless the animal is feeding very actively, that is unless it is swimming in a thick suspension of particles, the anterior third or half of the peritrophic membrane is empty of food. The posterior two-thirds or half of the peritrophic membrane is normally full of food undergoing digestion, the hindmost portion of the food being in the state of faeces ready for defaecation. The short rectum is free of faeces, its lumen being closed except for the periodic pumping in of water from the anus to the intestine. When defaecation takes place, which may occur several times in a minute, a short length of faeces is expelled in an instant through the rectum and anus.

The food within the peritrophic membrane is continuously moved to and fro by the antiperistalsis of the intestinal wall. Posteriorly the rapid, short antiperistaltic waves chum up the food. In the middle region of the intestine each of the slower and more powerful waves moves the food forwards a little way within the peritrophic membrane into its anterior empty region, and then the food at once moves back again after the wave has passed. In moving back again, the hindmost portion of the food which had been forced forwards starts to move first, followed immediately by the front portion. As the food moves forwards again, the anterior intestinal wall can be seen to be extended slightly by the liquid forced forward between peritrophic membrane and wall. As the food returns to its first position, the swollen anterior intestinal wall contracts. This contraction, whether muscular or merely elastic, would help the gut liquid and food to return back where they came from. The movements described must result in mixing the food and digestive enzymes.

The two caeca at the front of the intestine of Daphnia normally contain no food ; they contract together a fraction of a second after the food within the peritrophic membrane, forced forwards by intestinal antiperistalsis, starts to move back again. The contraction period of the caeca is thus the same as that of the middle intestinal antiperistalsis.^* The contraction of the caeca is apparently muscular and this too would assist the food to go back to its original position. Be this as it may, the liquid in the caeca is partly renewed at each contraction and subsequent expansion, and this must assist any digestive or absorptive function, as yet unknown, which the caeca may have.

This account of the movements of food in the intestine of Daphnia shows the important part played by the rhythmic movements of the intestine wall. A function of the oral as well as anal water intake may be to stretch the muscles of the gut wall so that they may contract rhythmically and effectively. In other Crustacea oral water intake would have an analogous function.

The rhythmic intake of water through the mouth may, in addition, have another function connected with the movement of food in the gut. In Cladocera, such as Daphnia, the food in the intestine, within the peritrophic membrane, can be seen to be gradually moved back towards the rectum, and periodically portions of the faeces are expelled through the rectum. What is it that moves the food backwards? When there are suspended food particles, algae or detritus, in the water, Daphnia feeds continuously, and the swallowing of food by the gullet must result in pushing backwards food that is already in the intestine within the peritrophic membrane. But when there is no longer any food to be swallowed, water is still taken in rhythmically and continuously through the mouth, while food that is already in the intestine is gradually moved on and periodically defaecated. It looks, then, as if it is the water which is continuously pumped in by the gullet that pushes the food backwards in the intestine within the peritrophic membrane. At the same time the swallowed water must distend the intestinal wall. Defaecation would take place when the rising pressure in the intestine caused a reflex release of the contracted rectum, whereupon the contraction of the stretched intestinal muscles, and perhaps also the elasticity of the intestinal wall, would expel a portion of faeces, ^† Perhaps in other Crustacea, too, water swallowed through the mouth moves food backwards along the gut until rising pressure, due in part also to anal water intake, leads to defaecation.

It is apparent that in order to contract rhythmically and effectively, the muscles of the intestine of Crustacea must be stretched by hydrostatic pressure. How is it, then, that the circular muscles of gullet and rectum, without being thus stretched, are able to pump water into the intestine, distending the muscles of the latter? The answer is that pharynx and rectum usually have radial dilator muscles (Weismann, 1874; Claus, 1876; Nowikoff, 1905; Humperdinck, 1924; Janisch, 1924; Ringel, 1924; Binder, 1932). As pointed out by Weismann (1874), the dilator muscles act as a suction pump, aspiring water to be forced into the intestine by the peristalsis or antiperistalsis respectively of the circular muscles of gullet or rectum. In aspiring the water, the dilators necessarily stretch the circular muscles and this enables them to pump the water inwards against pressure.

The action of the dilators of the gullet can easily be watched in living Daphnia and those of the rectum in Anostraca and in young Triops. In Chirocephalus diaphanus, although the gulps of water taken in by the rectum are small, rhythmic contractions of the dilator muscles are so strong that they move the body wall of the last abdominal segment. In Daphnia there are dilator muscles of the anus but not of the whole rectum (Binder, 1932). The action of these dilators in sucking water into the rectum may easily be watched under a low-powered microscope by putting Daphnia into a 5 % solution of urethane, when the animals can be made to He on their back. In Daphnia the rectal constrictor muscles are particularly well developed and are evidently strong enough to pump the aspired water into the intestine without themselves being stretched.

It is clear that most, or perhaps all, of the water taken in by prawns through the anus is expelled again through the same aperture at defaecation, or at pseudo-defaecation of water alone. In smaller Crustacea, with more frequent or even continuous rectal swallowing, water is certainly expelled again with the faeces, although we do not know what proportion this is of the water taken in. In some cases what appear to be astonishingly large quantities of water are pumped into the gut by the rectum when there are only infrequent defaecations that could get rid of the water. For instance, in a metanauphus of Triops cancriformis 324 rectal gulps of water were swallowed in $634$ min. without a defaecation; in an individual of Bythotrephes longimanus uninterrupted rectal swallowing of water at the rate of 92 gulps per min. (and oral drinking at the rate of 22 per minute) was watched for 5 min. without there being a single defaecation ; and in the small mysid Sirella armata, 8 mm. long, continuous rectal swallowing was watched for 15 min. with only one defaecation, the gulps being at the rate of 47 per minute.

The amount of water drunk through the mouth by Crustacea appears also to be large. What becomes of this water? In Leptodora there is sometimes antiperistalsis in the long midgut oesophagus and some of the water pumped into the stomach, from both mouth and anus, may then be vomited. But in Daphnia, Diaphanosoma and Limnadia I have never seen antiperistalsis in the gullet, and thus none of the water drunk continuously through the mouth is vomited. In most Crustacea other than Cladocera and Limnadia the gullet cannot be clearly observed. If the water taken in through the mouth is not usually vomited, is it got rid of through the anus? This cannot be its fate since the water coming in through the anus itself is, in many instances, apparently not all expelled by that path.

The water which we ourselves drink passes through the gut wall into the blood and is expelled from the body through the kidneys. In Crustacea the excretory organs, antennary or maxillary glands, must have a source of water too, and this water must enter the blood from either the gut or the gills. But is the very considerable quantity of water which is drunk by Crustacea through the mouth, with some at least of that swallowed by the rectum, largely passed through the gut wall and out of the body by the excretory organs? In favour of this is the following observation. If Daphnia is made to swim in a dilute solution of a dye, for instance bromo-thymol blue (Fox, 1948, p. 206) or nigrosin, after a few hours the dye can be seen to have accumulated in the liquid found in the anterior part of the intestine between peritrophic membrane and gut wall. The dye here is surprisingly concentrated and the process of accumulation is rapid. After Daphnia magna had been left overnight in a dilute solution of nigrosin, the concentration of dye in the lumen of the anterior part of the intestine was estimated. Animals were chosen which had no food in that part of the gut. The depth of colour of the dye in the gut of these animals, held lightly in a compressorium, was matched with that of the original solution in a flat-bottomed glass tube on the microscope stage (using a 2 in. objective), the height of solution in the tube being adjusted to give approximate equality of light absorption. The ratio of the height of solution in the tube to the internal diameter of the intestine showed that Daphnia had concentrated the dye about 250-fold. In nature the liquid between peritrophic membrane and gut wall of Daphnia is often bright green, the colour being due to a derivative of the chlorophyll of the algal food. The obvious explanation of this concentration of solutes in the gut lumen is that water is withdrawn through the gut wall from the solution. Perhaps the excretory organs of Crustacea require a particularly good flow of water and a further function of the more or less continuous oral drinking is to supply this water.

In the majority of small and transparent Crustacea water can be seen to be pumped into the alimentary canal through the anus by rhythmic antiperistaltic swallowing movements of the rectum. This anal drinking is continuous in small species and in the young of larger species, but occurs in intermittent bursts in the adults of larger species.

In prawns the intermittent anal intake of water acts as an enema, for it occurs only at the time of defaecation, which is preceded by one or two dozen rapid rectal gulps of water. The continuous anal intake of water by smaller Crustacea acts likewise as an enema, being continuous because of the more frequent defaecations, due to the higher metabolism and therefore greater food requirements of small animals. The water acts as an enema as in man, stretching the gut-wall muscles until they contract.

In prawns the rectal swallowing of water initiates and maintains intestinal antiperistalsis, which moves the swallowed water forwards in the intestine towards the thorax. A further function of the anal intake of water is thus to stretch the gut-wall muscles until they contract antiperistaltically. This is comparable with the initiation and maintenance of the heart beat in molluscs by hydrostatic pressure.

In the past it has been thought that the rectal swallowing of water by Cladocera is respiratory. This opinion was apparently strengthened by experiments showing that a deficiency of dissolved oxygen increases the rate of rectal swallowing movements. These experiments have not been confirmed, and other reasons are given which make a respiratory function unlikely.

Two further possible functions of the intake of water through the anus, namely the collection of salts necessary for osmoregulation, and a hydrostatic maintenance of body turgor, are discussed, tested and rejected.

Water can be seen to be swallowed more or less continuously through the mouth, by rhythmic peristaltic movements of the gullet, in the majority of small and transparent Crustacea.

A function of this oral drinking appears to be the same as that of anal drinking, namely to stretch the muscles of the gut wall. In Daphnia the antiperistaltic contractions of the midgut wall, which mix food and digestive enzymes, seem to be maintained by the hydrostatic pressure of water pumped into the gut by both gullet and rectum, defaecation occurring when this pressure rises to a certain value.

A second function of the rhythmic oral drinking by Daphnia and perhaps other Crustacea is to force the food in the midgut back towards the rectum.

The gullet and rectum of Crustacea have dilator muscles inserted into the exoskeleton. These muscles suck in water through mouth and anus, and by stretching the circular muscles they enable the latter to pump the water into the gut.

Much more water is taken into the gut of Crustacea than makes its exit at defaecation. Evidence is given that this water passes through the gut wall into the blood and out of the body by way of the excretory organs.

This work was begun in June 1948 when it was observed while examining metanauplius larvae of Apus, recently rediscovered in England (Fox, 1949), that water is continuously taken into the gut through the anus. Much of the work has been done in the Zoology Department of Bedford College, London, with material collected locally or sent from Plymouth, and considerable parts of the study were made at the Istituto di Idrobiologia at Pallanza in 1948, 1950 and 1951, and at the Station Zoologique at Villefranche-sur-mer in 1949. My grateful thanks are due to the Directors of these two laboratories, the late Prof. Edgardo Baldi and Dr G. Trégouboff, for their hospitality and help, and to Prof. Vittorio Tonolli at Pallanza for much willing assistance. Dr Sandro Ruffo took great trouble in collecting Palaemonetes at Verona and sent the prawns alive to Pallanza, and the late Dr A. Pacaud kindly arranged for Atyaephyra to be collected for me from a river in the Loiret Department in France. Penilia and Podon were studied at the Stazione Zoológica at Naples, and phyllopods were hatched out in London from dried mud derived, with the help of friends, from various parts of the world.