Studies On the Fate of Carbohydrates in the Rumen of the Sheep

In this present paper an account is given of experiments planned to investigate :

1. the rate of fermentation of specific carbohydrates in the rumen,

2. the products of fermentation in each case,

3. the fate of the products formed.

As regards rate of fermentation, it proved that the carbohydrates fell into three classes which are typified by (1) glucose, (2) galactose, (3) cellulose. The behaviour of the pH of the ingesta is characteristic for each group.

Sheep with permanent rumen fistulae or with both rumen and abomasal fistulae were used (Phillipson & Innes, 1939). Food was withheld overnight and samples were taken from the rumen before and at appropriate intervals after dosing. A standard dose of 100 g. was used except with cellulose which was given in the form of filter-paper pulp; of this a dose of 100 g. was too bulky and smaller quantities were given.

The same analytical methods were used as in the previous work (Phillipson, 1942). In many experiments large quantities of lactic acid accumulated in the ingesta, and as this acid is partially volatile, tests were made to find the error caused by this factor in estimating the volatile lower fatty acids. The distilling flask used consisted of a 300 c.c. Kjeldahl flask fitted with a double-bulb splash head; no lactate was found in the distillates and therefore it can be assumed that no error in the estimates for volatile acids occurred from contamination with lactic acid.

Glucose in the ingesta was estimated after suitable dilution by the copper reduction method described by King, Haslewood & Delory (1937) for blood. Sodium was estimated by the pyro-antimonate micro-titration method of Kramer and Gittleman described by Peters & Van Slyke (1932).

A few preliminary experiments in which the pH only was measured showed that after administering 100 g. of glucose or cane sugar a rapid fall in pH occurred during the following hour; fermentation was intense, judging by the quantity of gas that escaped from the cannula when the stopper was removed and by the frothing of the ingesta. No such reaction occurred after the administration of starch or cellulose; in the former case a slight but steady fall in pH was found which continued throughout the day, while with cellulose little change was seen. These observations were subsequently confirmed when full analyses were done.

The results can be divided into three groups, those in which intense fermentation occurred characterized by the accumulation of large quantities of lactic acid, those in which fermentation was less marked and in which no lactic acid or only small quantities were found and those in which fermentation was prolonged and in which no violent fluctuations occurred.

Intense fermentation was produced by dosing glucose, fructose and cane sugar; lactic acid accumulated quickly and was accompanied by a marked fall in pH; larger quaritities of lactic acid were found after dosing glucose and fructose than after dosing cane sugar, and it persisted for longer in the rumen. Volatile acids increased in all cases, but the rise was slower than that of lactic acid, the peak following that of the lactic acid which was greatest from $12$ to 1 hr. after dosing. The following example is typical of the results obtained from the glucose group.

The figures for volatile acids are given under two columns A and B. The former are the actual values expressed as grams of acetic acid per cent, while the latter are the values obtained after correcting for the fall in volatile acids that occurs if no dose is given. It was found previously that the quantity of volatile acids in the rumen gradually decreased when food was withheld (Phillipson, 1942). These experiments were commenced 24 hr. after the last feed, and from the 24th to the 48th hour the volatile acids steadily fall except for a slight rise that occurs at the 40th hour. On the basis of the average figures obtained. from two sheep, the following figures have been used to correct all values for volatile acids to a stable base line :

These corrections were added to the actual values found and the results obtained which are given under column B for all experiments indicate the true increase in volatile acid that occurred as a result of administration of the substance in question.

The second type of reaction was found after dosing galactose, lactose and maltose. No lactic acid was formed except with maltose when small quantities accumulated in one experiment. The fall in pH was slow compared with that which occurred when intense fermentation took place; the lowest point was reached from 4−5 hr. after administering galactose or lactose but earlier after dosing maltose. The rise in volatile acids, generally speaking, was reflected by the change in pH, but in some experiments increase in volatile acids persisted after the pH had commenced to rise again. This type of response is illustrated by the following experiment.

The initial fall in total volatile acids after dosing was found in many experiments and is presumably due to dilution of the ingesta although the water in which the sugar was dissolved was kept as small as possible; stimulation of the salivary flow as a result of dosing is also a possible explanation although the fact that no rise in pH occurred does not suggest that this was marked.

The last type of reaction was obtained only after dosing starch or cellulose. In both cases the fall in pH and the rise in volatile acids was slight but prolonged. In some experiments the starch was baked before administration in order to split up the granules but this did not produce a greater response.

These two experiments are quoted as they show the maximum change observed after dosing starch and cellulose respectively ; in both the rise in volatile acids is prolonged and the corrected figures under column B indicate that the fermentation of starch proceeds less slowly than the fermentation of, cellulose.

The results of the experiments with sugars agree with those obtained by McAnally (in Press) who studied the fermentation of carbohydrates by liquid rumen contents in vitro. She estimated the rate of fermentation by the evolution of gas and found that glucose was fermented quickly; appreciable quantities of gas accumulated after 10 min. incubation and steadily increased during the 90 min. period of observation. Cane sugar behaved in much the same way except that a slight initial lag occurred. Fermentation of maltose was slow in comparison but proceeded more quickly in the later stages, while little fermentation at all was obtained with lactose.

The fate of glucose was investigated in greater detail as it has been suggested that it forms the product of cellulose fermentation that can be stored in the body (Woodman, 1930). This involved the study of the rates of appearance and disappearance of lactic acid and of the volatile acids as both are produced when glucose is introduced into the rumen. The results are best described under the headings of the substance in question, and to do this it is necessary to split up the experimental results; each experiment in which glucose was dosed involved estimations of all three substances and the data are too bulky to discuss in any other way.

The disappearance of glucose from the rumen after dosing too g. was followed in four experiments; in one experiment glucose had disappeared in hr., and in the other three it had disappeared in from to to 2 hr. after administration. In these three experiments abomasal ingesta were taken as well ; in one, no trace of any reducing substance could be found in any one of the samples taken, while in the other two experiments only relatively small amounts were found in the abomasal ingesta. The details of two of these experiments are as follows:

It is obvious that the rapid disappearance of glucose from the rumen cannot be accounted for by passage to the abomasum. The quick appearance of large quantities of lactic acid and the increase of volatile acids which accompanies the disappearance of glucose indicate that fermentation is rapid.

The rate at which glucose disappears in vitro was also investigated. Samples of fluid obtained by filtering the ingesta through muslin fermented glucose rapidly provided excess substrate was not introduced, but the supernatant fluid obtained by centrifuging rumen ingesta possessed little fermentative power, an observation previously made by McAnally (in Press). If too much glucose was added, fermentation commenced rapidly but slowed down after 4 hr. or so ; this was due presumably to the accumulation of lactic and volatile acids with a corresponding drop in pH. In no case did fermentation proceed as rapidly as in vivo. In the experiment quoted below the substrate was present in excess and complete disappearance did not occur; the rate of fermentation was rapid during the first 4 hr., 0·296 g. % of glucose were fermented in 2 hr., 0·428 g. % in 4 hr., but only 0·496 g. % in 6 hr.

In two other experiments 0·91 and 0·62 g. % of glucose disappeared in $712$ hr. at 37°C., the quantities of lactate formed being 0·148 and 0·366 g. % respectively, while volatile acids expressed as acetic acid increased by 0·18 and 0·13 g. % respectively over the control flasks.

There can be no doubt that glucose is fermented by rumen ingesta and that lactic and volatile acids are produced ; the fact that fermentation is slower in vitro than in vivo is only to be expected, as no means of eliminating the products of fermentation are present. The quantity of lactic acid and volatile acids produced was small in comparison with the amount of glucose lost, but as no account was taken of the carbon lost as methane and carbon dioxide, quantitative results were not expected. The fate of glucose in the rumen is clear; it is rapidly fermented and little, if any, passes to the abomasum. It is, however, possible that some is directly absorbed from the rumen or during its passage through the omasum, and this aspect will be discussed later.

The rapid appearance and disappearance of lactic acid suggests that it is a transitory stage in the fermentation of glucose by bacteria of the rumen, especially as Woodman & Evans (1938) found that mixed cultures of cellulose-splitting organisms from the rumen of the sheep attacked lactate, the fermentation being characterized by the production of volatile fatty acids.

Solutions of sodium lactate were administered into the rumen, and the rate of disappearance together with the increase of volatile acids was estimated. Seven experiments were done and doses ranging up to 25 g. sodium lactate were administered; in all the experiments lactate disappeared from the rumen, in some cases within 2 hr. of dosing and in others within $412$ hr. The quantity of volatile acids produced was variable, irrespective of the size of the dose ; in only two experiments was there a marked rise ; in the other experiments the level of volatile acids remained more or less constant. An example of each type of result is given, both of which were obtained from the same sheep.

In the first experiment volatile acids fell slightly and the rise in pH suggests that this was due in part to an increase in the flow of saliva; although the dose was larger than that given in the other experiment, apparently little of the lactate was converted to volatile acid. On the other hand, when the smaller dose was given the increase in volatile acids was substantial.

In vitro experiments were also performed in which the disappearance of lactate was followed and the volatile acids formed were estimated. These experiments were subject to much variation. The supernatant fluid obtained after centrifuging the rumen ingesta had little activity upon lactate, but the activity of ingesta filtered through muslin exerted a variable effect. It was found, however, that if the amount of sodium lactate introduced was small, disappearance occurred within 5 to 6 hr.; in later experiments lactate was added to the rumen liquid in quantities not exceeding 0·2 g. %. The rate of disappearance under these conditions was comparable to that which occurred in vivo. The effect of introducing sheep saliva (filtered through a Berkfeld filter) into the fermentation flask as a buffer had no favourable effect upon the rate of disappearance of lactate when compared with flasks in which the same quantity of distilled water was added. The loss of lactic acid if expressed in 0·1 N % was approximately equal to the increase of total volatile acids expressed in the same way. The following experiment is typical :

In this experiment a flask containing 100 c.c. of rumen liquid, 10 c.c. of water and 1 c.c. of lactate solution was included; the result in this flask was almost exactly the same as that found in the flask containing saliva. The lactate flask lost 0·183 g-% lactic acid and gained 0·17 g. % volatile acids expressed as acetic acid over the control flask. These figures given in c.c. 0·1 N % show that the loss of lactate is accounted for by the gain in volatile acid; the lactic acid lost is equivalent to 20 c.c. 0·1 N % and the volatile acid gained to 28 c.c. 0·1 N %. The close correlation of the loss of lactate and gain in volatile acids suggests that the reaction proceeds molecule for molecule, and that the probable acid formed is acetic or propionic. Similar results were obtained in two other experiments.

As lactate may persist in the rumen for 4 hr., samples were taken from the abomasum in two experiments after dosing lactate into the rumen and in two experiments after dosing glucose into the rumen. Appreciable amounts of lactate appeared in the abomasum after dosing 36·3 g. sodium lactate, but only traces appeared after administering half this dose into the rumen. The quantity of lactic acid appearing in the abomasum after dosing 100 g. glucose into rumen also depended upon the concentration of lactic acid reached in the rumen. The details of these experiments are as follows:

The rate of disappearance of lactate in these sheep with double fistulae was slower than in the other animals, consequently the chances of lactate appearing in the abomasal ingesta were increased; even so, the quantity found was not large in comparison with the amount found in the rumen. In the latter two experiments the quantity of glucose passing to the abomasum was estimated, in the first no glucose appeared while in the second a little was found 15 min. after dosing but at no other time.

The fate of lactate therefore is similar to the fate of glucose except that it does not appear to be so rapidly fermented in vivo, consequently a larger proportion passes to the abomasum. The concentration of lactic acid in the abomasum never approaches that of the rumen, and as the volume of the abomasal contents is approximately one-tenth that of the rumen the loss of lactate from the rumen cannot be accounted for by passage to the abomasum. Lactic acid is not stable in the rumen ingesta and so its disappearance from the rumen can be attributed in part to fermentation but the possibility of direct absorption must still be considered.

It was found in experiments performed on a sheep with an abomasal fistula only that the quantity of volatile acids present in the abomasal ingesta was small and did not increase significantly after feeding as would be expected from the large quantities formed in the rumen. The animal was fed on bran and oats together with chaffed hay and straw, a diet known to increase the concentration of volatile acid in the rumen by 100% or over in from 4 to 8 hr. The following experiment illustrates the results obtained:

Additional proof that little volatile acid was present in the abomasal ingesta was obtained by estimating the active chloride; it was found that the greater part of the titratable acid could be accounted for by gastric secretion, and consequently that only small proportions of the titratable acidity, often less than those found in the above experiment, were due to organic acids. As the volatile acids are considered to be an end-product of carbohydrate fermentation in the rumen, and as the concentration of these acids present in the rumen is considerable, ten or more times the concentration found in the abomasum, it was decided to investigate the fate of acetic and butyric acids when introduced into the rumen.

The acids were given as the crystalline sodium salts, and a standard dose of 50 g. was used. Five experiments were done in which acetate was administered and two in which butyrate was given. In three experiments the sodium of the rumen and abomasal contents was also followed in order to see whether this behaved in the same manner as the acid radicle. Two control experiments were also performed in which it was found that an increased concentration of sodium occurred in the abomasal ingesta after adding 200 c.c. of water to the rumen, consequently the abomasal analyses for sodium after dosing sodium salts of the volatile acids are of no value. The results of these experiments are given in detail in Table 1.

It will be seen that in two of the experiments in which acetate was administered the results were disappointing; in Exp. 1 only a slight rise in volatile acids occurred after dosing, while in Exp. 3 the figures were extremely irregular. In the other experiments and in those in which butyrate was given, there was a marked rise of volatile acids Which in two cases did not reach its peak until 1 hr. after dosing; the concentration did not remain high but rapidly declined until by the sixth hour it had reached the region of the predosing level.

In two of the three experiments in which the sodium in the rumen was followed, the initial rise in sodium was proportional to the rise in volatile acids, and the subsequent decline in concentration followed that of the volatile acids reasonably closely, but in Exp. 4 the maximum concentration of sodium preceded that of the acetate and the subsequent decline was not maintained. The irregularity in this experiment may be due to the saliva which contains sodium in a concentration similar to that of blood, namely, in the region of 300 mg. %, consequently an increase in the inflow of saliva to the rumen would increase the concentration of sodium in the ingesta.

In no experiment did volatile acid reach the abomasum in any quantity; the slight increase in concentration that occurred in some experiments was insignificant.

These experiments established two main facts; the first is that the increased con4 centration of volatile acids produced in the rumen by administering acetate or butyrate is not maintained and the fall in concentration is rapid compared to that found during fasting (Phillipson, 1942). The second is that only traces of volatile acids are present in the abomasal contents and little or no increase is produced either by feeding or administering acetate or butyrate into the rumen. The possibility that the volatile acids are unstable in rumen ingesta had still to be investigated, and to do this five experiments were performed in vitro; in the first, 100 c.c. of liquid rumen ingesta was placed in each of four flasks; to one was added 10 c.c. of water and to the others were added 10 c.c. of a known solution of either sodium acetate or propionate or butyrate. Precautions were taken to keep the rumen ingesta warm. The flasks were incubated for $812$ hr. at 38° C., and at the end of this time the total volatile acids in each flask was determined by steam distillation. The result of this experiment showed that no loss of volatile acid occurred.

Assuming that the increase in volatile acids in each of the three experimental flasks was the same as that occurring in the control flask, then the figures in the left-hand column represent the quantity of volatile acid that should have been present after incubation. The actual quantities determined by steam distillation are given in the right-hand column, and in each flask the amount present was greater than expected; this suggests that fermentation of the existing substrate present in the rumen ingesta was slightly greater in the flasks to which volatile acid was added than in the control flask; there is no evidence that the acids added had been attacked by the bacteria.

In the second experiment four flasks were put up, one as control and the others as experimental flasks; to each of the latter sodium acetate, sodium propionate or sodium butyrate solution was added. A sample of 10 c.c. was taken from each flask before they were placed in the incubator and the total amount of volatile acids present was estimated. The contents of each flask was covered with liquid paraffin to prevent evaporation and further samples were taken after $712$⁠, 24 and 48 hr. incubation. The total volatile acidity was expressed as a percentage and by subtracting the quantity of volatile acids present in the control flask at each period of incubation from the quantity present in the other flasks at the same period, the amount of volatile acid added was obtained. These values are given in the following table:

It will be seen that during the first $712$ hr. the amount of acetate and butyrate apparently increased and propionate slightly decreased, but the changes either way were not marked. The volatile acids were only significantly reduced after 24 hr. incubation, while during the last period the rate of decrease was accelerated to such an extent that a minus figure was obtained with acetate and butyrate, especially the former.

The instability of the volatile acids on prolonged incubation in rumen ingesta in no way invalidates the fact that they were stable for $712$ hr. ; the succeeding decrease, which became more rapid with time, indicates that a rapid multiplication of acid-splitting bacteria was occurring. The initial period in which the volatile acids are stable indicates that acid-splitting organisms are normally numerically insignificant in the rumen and only gain preponderance under in vitro conditions.

The results of these experiments were confirmed by three further trials.-Approximately normal solutions of sodium acetate, propionate and butyrate were made up and the strengths of their solutions were determined by steam distillation. Different amounts were used in each experiment, but in no case was the concentration of total volatile acids in the flasks raised above 150 c.c. 0·1 N %, a figure that is within thé range found normally in the rumen ingesta. A sample was taken from each flask after it was put up and the quantity of volatile acids present was determined; in addition, in two of the experiments a sample of the rumen liquid, equal in quantity to that put into the flasks, was distilled; thus the total quantity of volatile acids introduced was estimated twice. The contents of the flasks were covered by liquid paraffin and incubated for 8–$812$ hr.; a further sample was taken and the volatile acids present after incubation were estimated. The results are expressed as the total volatile acids present in the flask in c.c. 0·1 N% assuming that none of the contents were removed during sampling. The full results of the three experiments are given in Table 2.

It will be seen that the results are fairly regular and no significant loss occurred in any experiment. The variations shown in the last column are better expressed as percentages of the amount of acid originally added, as the concentration was different in each Experiment. Percentage figures are given in the following table together with the corresponding figures for the first two experiments:

The figures deviate from + 20 % to – 20 %, and with each salt plus and minus figures of nearly equal amplitude occur. These deviations are too wide to be accounted for by error in distillation and must be due to slightly different rates of fermentation in the experimental flasks. Even if the greatest loss, namely 20%, is considered alone, this is insufficient to account for the rapid disappearance of acid from the rumen. On the basis of these experiments there is no reason to suppose that acetate, propionate and butyrate are readily broken down by the bacteria of the rumen as are lactate and glucose; in fact, they appear to be sufficiently stable in rumen ingesta to warrant the conclusion that the rapid disappearance from the rumen of volatile acid added in excess of the quantity present is not due to bacterial activity.

The fate of volatile acids appears to be that they are absorbed before they reach the abomasum; this conclusion rests upon the circumstantial evidence that they rapidly disappear from the rumen when introduced in quantités in excess of that already present, that they do not pass to the abomasum and that they are stable for 8 hr. when incubated in vitro in rumen ingesta. If the conception that absorption can occur readily in the rumen or omasum is accepted, then the fact that the volatile acids are always present in the rumen is due not so much to the fact that they are end-products of fermentation but to the fact that they are continually being produced. The rate of production during fasting, when the level of volatile acids is steadily falling, is sufficient therefore only to prevent a rapid disappearance of volatile acid from the rumen and not to maintain a steady level. Support is given to this conception by an experiment in which 25 g. lactic acid was introduced into the rumen. The pH of the ingesta fell abruptly to 4·7 and the volatile acids disappeared within 15, min. of dosing; this suggests that the fall in pH either killed or temporarily inactivated the cellulose-splitting bacteria so that the supply of volatile acid failed and the acids already present were rapidly absorbed.

The difference in the rate of fermentation of sugars in the rumen can be correlated with sugars occurring naturally in foodstuffs. Cane sugar is present in mangolds while glucose and fructose are primary cleavage products. The smaller quantity of lactic acid that accumulated when cane sugar was introduced into the rumen, compared with the amount that was found after dosing glucose and fructose, suggests that this is due to a slight lag while the disaccharide is being hydrolysed to produce glucose and fructose. Lactose and galactose do not occur naturally in the rumen; in suckling lambs milk passes directly to the abomasum and does not enter the rumen (Phillipson, 1939), consequently the slow rate of fermentation is not surprising. It is possible that lactic acid is also a stage in the fermentation of these sugars, but as the process is slow it is further broken down as fast as it is formed and never accumulates. As lactate may appear during the fermentation of maltose which proceeds at a greater speed than that of lactose, this theory has some support.

The ciliates of the rumen must be considered in relation to the fermentation of starch. It is easy to demonstrate under the microscope that these organisms ingest starch granules; Becker (1932) stated that lambs which have been rendered free of ciliates, by fasting followed by a dose of 2 % copper sulphate, became rather distended with gas in the rumen after eating food containing maize meal; this observation suggests that the ciliates assist the host by ingesting starch, so protecting it from bacterial fermentation while the ciliates themselves, in which the starch is said to be converted to glycogen (Trier, 1926), may be finally digested by the sheep when they reach the abomasum. The sheep used in these experiments all contained a flourishing fauna in the rumen, and the slight reaction observed after dosing starch may be due partly to the ciliates, although as starch is a polysaccharide it cannot be expected to be fermented so readily as sugar. Steinhauf (1921) found that the abomasal contents of sheep developed a distinct blue staining reaction with iodine after eating crushed oats, while Schalk & Amadon (1928) observed in cows with rumen fistulae that foods such as maize meal when eaten passed to the reticulum and that some passed directly through the omasum to the abomasum without entering the rumen ; the action of the reticulum, which contracts briskly approximately once a minute and throws its contents backwards to the rumen, does not allow foods such as maize meal to remain long enough in the region of the omasal orifice.for more than a portion of the ingesta to take this direct route. Starch, therefore, may be digested in three ways in the ruminant: some may pass directly to the abomasum and undergo normal digestion as in other animals, some is ingested by the ciliates in the rumen and may be conveyed to the host by these organisms in one form or another, while some is subjected to bacterial fermentation. The proportion that is accounted for by each of these three routes remains a subject for research.

There is ample evidence that the fermentation of cellulose is slow; no one has yet demonstrated the presence of a cellulose-splitting enzyme in any of the digestive secretions of any mammal, while the evidence concerning the ciliates of the rumen all points to the fact that these organisms cannot utilize cellulose in any way (Mangold, 1933); digestion of cellulose in the ruminant therefore must be due to bacteria alone. The observations of Ritzman & Benedict (1938), who found in cattle that the production of methane did not cease until 4 days after the last meal when the animals were fed hay, suggests that fermentation of cellulose must persist in the rumen during this period. McAnally (1942) found that filter paper when enclosed in silk bags required 3 − 4 days for complete digestion to occur in the rumen, although the presence of the silk barrier probably delayed the onset of fermentation ; the prolonged reaction obtained in these experiments after dosing filter-paper pulp into the rumen is in agreement with these findings.

The rapid fermentation of glucose in the rumen rules out the possibility of glucose as such being the principal end product of Cellulose fermentation; however, glucose might be made available to the host for absorption if, as Baker (1939) suggests, the iodophilic bacteria, besides hydrolysing cellulose, deposit within their protoplasm a glycogen-like substance which represents the condensation of soluble sugars obtained from cellulose. This may be later digested by intestinal amylases and so reach the host as glucose. The hypothesis originated by Pochon (1935) that glucose is formed in the abomasum from cellulose cannot be true, as no glucose can be found in the abomasal ingesta.

The failure to account for the fate of the volatile acids either by passage to the abomasum or by bacterial decomposition suggests that these acids are directly absorbed either in the rumen and reticulum or during their passage through the omasum. If this is possible then it is also possible that the disappearance of glucose and lactate may also be due in part to direct absorption. The rumen and reticulum are generally stated, on anatomical grounds, to be merely storage and fermentation chambers; the epithelium lining these organs is stratified and squamous, and until recently this has been sufficient reason to deter curiosity regarding absorption from these organs. The omasum is an organ particularly suitable for absorption, as the laminae offer a large surface area which in the cow is equal to that presented by the whole small intestine (Colin, 1886). Other evidence is available that absorption in the non-glandular part of the stomach does occur; thus Aggazzotti (1910) found that water and sugars in solution were absorbed from the omasum but to no great extent from the reticulum or abomasum. Trautmann (1933) proved that drugs such as pilocarpine and atropine were rapidly absorbed through the mucosa of the rumen while Lenkeit & Columbus (1935) found that after dosing potassium iodide to sheep and lambs no difference in the rate of excretion of iodine in the urine occurred, although with adult sheep the dose passed to the rumen and in the lambs it passed directly to the abomasum.

These observations, together with the evidence presented in this paper, although it is admittedly negative evidence, are strong enough to warrant a thorough investigation into the question of absorption from the ruminant stomach and such a programme of research is planned.

1. Glucose, fructose and cane sugar undergo rapid fermentation in the rumen of the sheep and pass through the stage of lactic acid to volatile lower fatty acids. Maltose, lactose and galactose are fermented less rapidly and lactic acid, if it is a stage in the fermentation, does not accumulate in the ingesta. Starch and cellulose are fermented slowly and the production of volatile acids is prolonged.

2. The rapid disappearance of glucose from the rumen cannot be accounted for by passage to the abomasum.

3. Lactic acid is not stable in the rumen; when it is present in quantity, some appears in the abomasum.

4. Volatile acids are stable in the ingesta of the rumen; they do not pass to the abomasum in any quantity and appear to be absorbed in the rumen or omasum.

5. The progress of digestion of carbohydrates in the stomach of the sheep is discussed.

The first part of this work was done while one of us (A. T. P.) was holding the ‘Miss Aleen Cust’ Scholarship awarded by the Royal College of Veterinary Surgeons. The authors wish to express their thanks to Sir Joseph Barcroft, to Prof. H. A. Harris and to Prof. T. Dalling for their help and encouragement, to Mr E. I. McDougall who assisted in some of the experiments and to Mr E. Turner for his technical help.