The integrated processing response of voles to fibre content of natural diets

Authors

  • M. Young Owl,

    1. Department of Ecology, Ethology and Evolution, University of Illinois, Shelford Vivarium, 606 East Healey Street, Champaign, IL 61820, USA
    Search for more papers by this author
    • Present address: Departments of Anthropology and Biological Sciences, California State University, Long Beach, CA 90840, USA

  • G. O. Batzli

    1. Department of Ecology, Ethology and Evolution, University of Illinois, Shelford Vivarium, 606 East Healey Street, Champaign, IL 61820, USA
    Search for more papers by this author
    • For correspondence.


Abstract

1. Two species of voles were fed high fibre (barnyard grass) and low fibre (alfalfa) diets to test the integrated processing response (IPR) hypothesis. This hypothesis states that many herbivores are able to maintain their required intake of digestible nutrients and energy on diets with very different fibre content because of compensatory changes in intake of food, size of gastro-intestinal (GI) tract, passage rates of fibre and absorptive capacity of the GI tract.

2. As predicted by the IPR hypothesis, each species of vole maintained a similar intake of digestible dry matter on the two different diets. Both species also had greater intake, larger GI size, shorter mean retention times and greater GI mass (an indicator of epithelial mass and absorptive capacity) when fed grass than when fed alfalfa.

3. The two species differed in that meadow voles, the more active species, had greater total intake and obtained a greater amount of digestible dry matter from either diet than did prairie voles. Meadow voles also consume more grass in the field than do prairie voles, and they digested grass better than did prairie voles. Prairie voles, which consume more dicots in the field, digested alfalfa better than did meadow voles.

4. Meadow voles had longer GI tracts, particularly small intestines, than did prairie voles, which may be linked to their greater ability to digest grass. However, meadow voles did not have larger caeca than prairie voles, even though caecal size increased on grass diets for both species. The GI size of prairie voles fed grass increased more than did the GI size of meadow voles, and this may have enabled prairie voles to utilize a grass diet, though they prefer to eat dicots. Greater selection of leaves, which have less fibre than stems, and longer mean retention times of food may account for better digestion of alfalfa by prairie voles.

Introduction

Arvicoline rodents (lemmings and voles) live in a wide variety of habitats and eat a wide variety of herbaceous plants (Tamarin 1985; Stenseth & Ims 1993). Preference for particular plant parts or species depends on their quality, which is positively related to their content of digestible energy and nutrients and negatively related to their content of toxic secondary compounds and poorly digestible fibre, and on the digestive capabilities of the herbivore (Batzli 1985). Because plant availability and condition vary in space and time, the diet of an individual herbivore may sometimes be dominated by highly digestible items (low fibre content), such as seeds and dicot leaves, and other times by poorly digestible items (high fibre content), such as mature shoots of monocots (Batzli 1985).

Increasing fibre dilutes the concentration of digestible nutrients and energy in a diet. Yet, arvicoline rodents manage to obtain constant amounts of digestible energy and nutrients from diets with very different fibre content (Batzli & Cole 1979). Voles increase their intake of food to compensate for poor digestibility (high fibre). However, increased food intake leads to more rapid passage of food through the gastro-intestinal (GI) tract and a shorter mean retention time (MRT), which further reduces digestibility (Sibley 1981; Demment & van Soest 1985). Thus, some additional adjustment must occur to maintain a constant intake of digestible dry matter, and two such adjustments do occur in response to increased fibre in the diet or to increased energy demand. The size of the GI tract increases, which lengthens the MRT for a given level of intake, and the mass of the GI tract increases (Gross, Wang & Wunder 1985; Hammond & Wunder 1991). The latter response occurs in both total mass and mass per unit length of GI tract and largely consists of increased intestinal epithelium, which in turn increases the rate of nutrient absorption (Ferraris, Lee & Diamond 1989; Derting & Bogue 1993).

The simultaneous adjustment of dietary intake, size of GI tract, MRT and epithelial mass of the intestine in response to increased dietary fibre was named the integrated processing response (IPR) by Batzli, Broussard & Oliver (1994). At least some of these adjustments occur in a wide variety of herbivorous birds and mammals (Pulliainen & Tunkkari 1983; Gross et al. 1985; Brugger 1991 and references cited therein), and recent evidence suggests that they may even occur in grasshoppers (Yang & Joern 1994). This widespread occurrence led Batzli et al. (1994) to propose the IPR hypothesis, which states that many herbivorous animals are able to maintain their required intake of digestible dry matter or energy on diets with very different fibre content because of compensatory changes in food intake, GI size, MRT of the diet and absorptive capacity (epithelial mass) of the GI tract.

Several experimental studies on voles support the IPR hypothesis. Gross et al. (1985) first reported that prairie voles, Microtus ochrogaster (Wagner), fed rat chow diluted with cellulose (39·8% acid detergent fibre or ADF) for an 18 day acclimation period had greater food intake, longer caeca and heavier intestines than voles fed undiluted rat chow (7·8% ADF). Hammond & Wunder (1991) conducted a similar experiment but gathered a wider variety of data on prairie voles fed either a high fibre laboratory diet (22·5% ADF) or a low fibre laboratory diet (8·5%). They found lower digestibility but higher food intake, a longer GI tract, greater GI dry mass, greater wet mass of GI contents and similar digestible energy intake for voles fed the high fibre diet. In a separate experiment, but using the same diets, Hammond (1989) found faster passage rates (shorter MRT) for the high fibre diet. Batzli et al. (1994) acclimated two species of voles for 21 days to foods actually eaten in the field, either grasses with high fibre (27. 2% ADF) or alfalfa with low fibre (12·5% ADF), and then fed them the same or opposite diet. They reported that meadow voles, Microtus pennsylvanicus (Ord), had higher intake of digestible dry matter than prairie voles, but both species had greater intake when fed grasses than when fed alfalfa. Both intake and digestibility of grasses were greater for voles acclimated to grasses compared to voles acclimated to alfalfa, but only the intake of alfalfa was greater for voles acclimated to alfalfa. The net result was that both species obtained more digestible dry matter on diets to which they had been acclimated (10–13% improvement on alfalfa and 20–27% improvement on grass), which indicated the likely adaptive advantage of the IPR. Changes in GI size were not as expected, probably because animals were put on the non-acclimation diet for three days at the end of the experiment so that intake and digestibility could be compared to the acclimation diet. During that time the GI tracts may have begun to re-acclimate to the new diet. Although these studies lend support to the IPR hypothesis, they were not definitive because none of them examined all aspects of the IPR on the same animals, and the only study to use natural diets did not examine GI tracts until diets had been switched for three days.

The experiment reported here provides a more rigorous test of the IPR hypothesis by feeding natural foods to two species of voles that have different nutritional characteristics and by measuring all aspects of the IPR on the same individuals. Meadow voles have a greater intake of digestible dry matter, presumably because they are more active and have greater metabolic demand, and take more monocots (high fibre) in their diet compared to prairie voles of the same size (Batzli et al. 1994; Haken & Batzli 1996). The IPR hypothesis predicts that both species should show increased intake, larger GI size, higher passage rates (shorter MRT) and greater absorptive capacity (increased dry mass of GI tract) when fed higher fibre diets, but the daily intake of digestible dry matter within a species should be similar on low fibre and high fibre diets. Because of their greater intake of food, meadow voles were also expected to have larger GI size or higher passage rates, or both, compared to prairie voles fed the same diets.

Methods

For this experiment the progeny of two species of voles, prairie voles and meadow voles, were used. They were collected at the University of Illinois Ecological Research Area 5 km NE of Urbana, Illinois. All captive animals were maintained in polycarbonate shoe-box cages with corn-cob bedding and provided with rabbit chow and water ad libitum when not on experimental diets. Young were weaned from their parents at 20–25 days, then kept in sibling groups of the same sex until assigned to an experiment, after which they were housed individually. The experimental design was a 2 × 2 factorial with 10 replicates, using species and diet type as the two factors. For each species, five male and five female young adults (30–40 g body mass) were randomly chosen to receive one of two diets, except that no siblings were assigned to the same experimental group. One group received a high fibre, natural food (barnyard grass, Echinocloa crusgalli (L.)) ad libitum, the other a low fibre, natural food (alfalfa, Medicago sativa L.) ad libitum. After 3 weeks acclimation to the experimental diet, individuals were placed in metabolic cages for 3 days to measure apparent digestibility and mean retention times for the diets. The four experimental groups were run sequentially, rather than simultaneously, for logistic reasons.

Food plants were harvested from local fields in Champaign County, Illinois during September and October and oven dried at 60 °C before feeding to the voles. Previous work had shown that dried plants maintained their palatability (Batzli & Cole 1979). Only the top branches of alfalfa were harvested to avoid basal stems, which can have very high fibre content, but whole shoots of barnyard grass were used. None of the plants was flowering. To determine if voles selected either leaves or stems, samples of the food as fed and of the uneaten food (orts) were analyzed for ADF content using the method of Goering & van Soest (1970). Higher or lower fibre content in the orts compared to the diet as fed would indicate selection of plant parts with lower or higher fibre content, respectively. Because some values for proportion of ADF were below 0·3, arcsine tranformations of the data were used before conducting statistical tests on ADF content.

During the digestibility trials, the voles were fed marked fibre each morning. After consumption of the marker, faeces were collected every hour for the next 15 h, and again at 24 h before the next day's trial. At the end of the digestibility trials, all voles were weighed, humanely killed, placed in sealed plastic bags, and frozen until the GI tracts could be removed and measured. All orts and faeces were collected, separated, dried at 60 °C to constant weight. The apparent digestibility was estimated based on total collections for the three day period as: (food intake – faeces)/food intake.

To mark the fibre chromium oxide was mordanted on cell walls recovered from the faecal pellets of voles that had been fed the appropriate diet (barnyard grass or alfalfa) for over 24 h. This was done so that the sizes of the marked particles would mimic those that normally passed through the GI tract of the voles. The faeces were dried, ground with a mortar and pestle, boiled in a solution of 3% sodium lauryl sulphate for 1 h to remove the soluble materials and thoroughly rinsed with water and acetone over bolting silk. Following the methods of Uden, Colucci & van Soest (1980), the cell walls were then heated at 100 °C for 24 h in a sodium dichromate (Na2Cr2O7) solution containing an amount of Cr equal to 12–14% of the dry weight of the sample. All voles, except one, readily consumed a dose of 0·15 g of marked cell walls when mixed with apple juice.

The amount of chromium in the faecal samples collected from the experimental voles was measured using the method of Williams, David & Iismaa (1962), slightly modified by diluting the acid digestion with distilled water to only 25 ml, which prevented overdilution of chromium in small amounts of vole faeces (Williams et al. worked with ruminants). The chromium content of each sample was then measured by atomic absorption spectrophotometry. Mean retention times were calculated using the equation of Warner (1981):

inline image

where mi is the amount of Cr (concentration × dry mass) in an hourly sample of faeces and ti is the time of collection of the sample (midpoint of the collection period). Recovery of the Cr occurred sporadically because of fluctuations in faecal output and concentration of Cr in the faeces (see Results). Therefore, the MRTs were calculated using only the results for a single day for each vole, the first day that produced a recovery curve with a clear peak.

The contents of the GI tract were emptied and dried at 60 °C to constant weight. Estimates of GI size were made for each compartment (stomach, small intestine, caecum and colon) by measuring length and three widths (w) at 25%, 50% and 75% of the compartment length, taking care not to stretch the tissue. The volumes (V) were calculated, based upon lengths (l) and average radii (ri = w/2), assuming each compartment to be cylindrical (V = πri2l). Because the irregular shape of GI compartments introduces unknown errors into the calculations of volume, the gross surface area of each compartment was also directly measured, assuming that surface area should be directly related to volume. To do this the compartments were opened along one side, slit at other locations as necessary, flattened onto a piece of newsprint and outlined as described in Young Owl (1994). The outlines were then traced onto graph paper with 1 mm squares and the number of squares encompassed by each compartment counted. Finally, all GI compartments were dried at 60 °C to constant weight. The GI tract of one prairie vole fed barnyard grass diet was accidentally destroyed, which reduced the sample size to nine for that group.

Most voles maintained their body mass when fed either diet, and for three of the experimental groups, all animals lost less than 10% of their body mass during the acclimation period. However, a temperature excursion occurred during the second week of acclimation of the meadow voles to alfalfa. The room temperature peaked at 35 °C and did not return to normal room temperature (23–25 °C) for 3 days. Two of the animals died before the experiment ended. Another three animals survived but lost more than 20% of their body mass and appeared emaciated, so the data from these animals were not used. The five remaining meadow voles lost less mass, behaved normally, and had normal body fat ( = 7·9%) compared to the other experimental groups ( = 6·9–10·7%). It was therefore assumed that they had recovered fully from the temperature stress, and their subsequent performance (intake and digestion) on alfalfa supported this assumption. All animals used in the analyses maintained their body mass during the 3-day trials for digestibility and passage rates.

Although the body mass of voles assigned to experimental groups was originally restricted to 30–40 g, body sizes differed somewhat by the end of the experiment (three-way ANOVA for diet, species and sex). Males were slightly heavier than females ( = 35·1g and 33·5 g, respectively; P = 0·04) and prairie voles were heavier than meadow voles when fed alfalfa ( = 36·3 g and 29·8 g, respectively) but not when fed grass ( = 34·7 g and 34·5 g, respectively; P = 0·003 for interaction). Males also had greater head and body length than females when fed alfalfa ( = 117 mm and 110 mm, respectively) but not when fed grass ( = 112 and 113, respectively; P = 0·05 for interaction). Because of these differences in size and the likely effects of size on most of the measurements of food processing and of the GI tract, mean values for the adjusted data were calculated in addition to means for the unadjusted data. Adjustments were based on head and body length or live body mass, depending on which covaried most closely with the measurement (intake/live body mass, GI length/head and body length, GI surface area/head and body length squared, GI volume/live body mass, GI dry mass/live body mass, GI contents/live body mass). Both unadjusted and adjusted data met the criteria for parametric statistical analysis.

Three-way ANCOVA or ANOVA were first conducted on unadjusted data with diet, species and sex as the main factors and the appropriate measure of body size as the covariate. ANCOVA was used for data on intake, digestible intake and the length, surface area, volume, dry mass and contents of the GI tract. Sex influenced only three variables: females had greater stomach volume (P = 0·01), greater stomach dry mass (P = 0·0006) and greater relative stomach mass (dry mass/length, P = 0·03) than did males. In no case did sex interact with diet or species, which indicated that sex did not influence experimental results, so both sexes were included in two-way ANCOVA or ANOVA of the unadjusted data to test for significant effects of diets and species. Data summaries for unadjusted data and adjusted data showed the same major trends. However, summaries based on the adjusted data are presented, except for those measurements without covariates, because they more closely reflected the statistical results from ANCOVA than did the summaries of unadjusted values.

Results

Prairie voles selected less fibrous plant parts, particularly for alfalfa, but the meadow voles did not. Barnyard grass had a mean (± 1 SE) ADF content of 33·0 ± 1·1% for the whole shoot, 25·6 ± 0·2% for leaves and 34·6 ± 1·3% for stems (n = 6 in all cases). The ADF content of orts for prairie voles was 34·1 ± 0·6% (n = 5), which was similar to the grass as fed (t = 1·09, df = 9, P = 0·31) and indicated little selection for leaves. For the meadow voles, the ADF content of orts was 32·5 ± 0·8% (n = 5), which even more closely matched the original value for whole shoots (t = 0·30, df = 9, P = 0·77) and indicated no selection of plant parts. Alfalfa as fed had an ADF content of 22·5 ± 0·4% for the whole shoot, 16·8 ± 0·5% for leaves and 32·6 ± 0·8% for stems (n = 6 in all cases). Orts for the prairie voles contained 26·2 ± 1·0% ADF (n = 5), 16·4% more fibre than in the alfalfa as fed (t = 3·89, df = 9, P = 0·004), indicating significant selection for leaves, whereas orts for the meadow voles contained only 21·9 ± 1·0% ADF (n = 5), which closely matched the alfalfa as fed (t = 0·58, df = 9, P = 0·57).

Intake by the two species of voles showed similar responses to higher fibre in their food, but their overall consumption levels and abilities to digest the diets differed. The meadow voles consumed more food than the prairie voles, whether they were consuming grass or alfalfa; 49% more on grass and 47% more on alfalfa (Fig. 1a, two-way ANCOVA, P < 0·0001 for species). Within each species, those animals fed barnyard grass consumed more food than those fed alfalfa; prairie voles ate 23% more grass and meadow voles consumed 24% more grass (1Fig. 1a; two-way ANCOVA, P = 0·0002 for diet, P = 0·06 for interaction of species and diet). Apparent digestibility was much greater for alfalfa (53·2 ± 1·4% for grass and 69·0 ± 1·4% for alfalfa) for both species of voles (1Fig. 1b, two-way ANOVA, P<0·0001 for diet, P=0·33 for species and P<0·0001 for the interaction). The interaction occurred because meadow voles digested the grass much better than did prairie voles, whereas the prairie voles digested alfalfa slightly better than did the meadow voles. As expected from the results for intake and digestibility, the meadow voles digested a greater amount of dry matter (56% more) than did prairie voles, but there were no significant differences between diets (1Fig. 1c, two-way ANCOVA, P<0·0002 for species, P=0·43 for diet, P=0·007 for interaction). The interaction represents a higher digestible intake by the prairie voles fed alfalfa compared to those fed grass but a higher digestible intake by the meadow voles fed grass compared to those fed alfalfa.

Figure 1.

. Daily intake of dry matter (a), dry matter digestibility (b), daily intake of digestible dry matter (c) and total dry mass of gastro-intestinal contents (d) for prairie voles (MO) and meadow voles (MP) fed grass (hatched) or alfalfa (open). Bars show means and vertical lines show 95% confidence limits, n = 10 for all experimental groups, except meadow voles fed alfalfa (n = 5). Values for intake, digestible intake and gastro-intestinal contents adjusted for live body mass (gbm).

The total GI tracts of prairie voles and meadow voles also responded similarly to the fibre content of the diets, although individual compartments differed. The prairie voles generally had larger stomachs than did meadow voles, as shown by the length, surface area and volume, but not by dry mass (Table 1, note two-way ANCOVA species effects). The prairie voles fed grass had longer stomachs with greater surface area, volume and dry mass than did those fed alfalfa, but the meadow voles showed none of these trends (Table 1, note two–way ANCOVA interactions). Changes in mass and length of stomachs were similar so that the mass per unit length showed no significant response to different diets for either species.

Table 1.  . Adjusted means ± SE for size of gastro-intestinal compartments of two species of voles when fed grass or alfalfa. Sample sizes given in parentheses. For two-way ANCOVA: S, species effect; D, diet effect; I, interaction Thumbnail image of

The meadow voles had longer small intestines than did the prairie voles, but the surface area and volume did not differ between the species because the small intestines of prairie voles had slightly larger diameters (Table 1). The surface area and volume of small intestines increased for both voles when fed grass, but length did not. Because dry mass increased, the dry mass per unit length of small intestine also increased for voles fed grass (Table 1).

Caecal size (length, surface area and dry mass) increased for voles fed barnyard grass and differed little between the species. However, caecal volume increased only for prairie voles fed grass (Table 1), and the dry mass per unit length was consistently greater for the prairie voles than for meadow voles. Colonic length, surface area and volume were all greater for prairie voles than for meadow voles but did not differ between diets (Table 1). Dry mass per unit length of colon was greater in the voles that were fed barnyard grass, but did not differ between species.

The net result of all the differences in the GI compartments (total GI tract) can be calculated as the sum of all the compartments. Meadow voles had longer GI tracts than prairie voles, but diet had little effect on the total length of either species (2Fig. 2a, two-way ANCOVA, P=0·01 for species, P=0·39 for diet, P=0·87 for interaction). As expected, the total surface area and total volume were highly correlated (r = 0·93). Both surface area and volume increased with the grass diet (the prairie voles showed a greater response in volume than did the meadow voles), but the differences between species were not significant (2Fig. 2b; two-way ANCOVA for surface area, P < 0·0001 for diet, P = 0·09 for species, P = 0·12 for interaction; two-way ANCOVA for volume, P < 0·0001 for diet, P = 0·17 for species, P = 0·04 for interaction). Total dry mass (Fig. 2c) and total dry mass per unit length (Fig. 2d) increased for both species of voles when fed grass (two-way ANCOVA, P = 0·003 and P = 0·001, respectively), but differences between species (P = 0·47 and P = 0·30, respectively) and interactions (P = 0·16 and P = 0·42, respectively) were not significant.

Figure 2.

. Total length of gastro-intestinal tracts (a), total surface area (S) and volume (V) of gastro-intestinal tracts (b), total mass of gastro-intestinal tracts (c) and mass per unit length of gastro-intestinal tracts (d) for prairie voles (MO) and meadow voles (MP) fed grass (hatched) or alfalfa (open). Bars show means and vertical lines show 95% confidence limits, n = 9 for prairie voles fed grass, n = 10 for prairie voles fed alfalfa and meadow voles fed grass and n = 5 for meadow voles fed alfalfa. Values for total length, total surface area, total volume and total mass of gastro-intestinal tracts adjusted for head and body length (cmbl), head and body length squared (cmbl2) or live body mass (gbm).

The trends in GI contents of the two species were similar to those for the total GI volume. Prairie voles’ GI contents were 109% greater when fed grass than when fed alfalfa, whereas meadow voles had similar GI contents on both diets (1Fig. 1d, two-way ANCOVA, P<0·0001 for diet, P= 0·34 for species, P < 0·0001 for interaction).

The concentration and total amount of Cr in the faecal collections often showed more than one increase and decrease during a collection period, apparently because of coprophagy (Lee & Houston 1993a), and the temporal patterns often differed considerably among different individuals on the same diet and among consecutive daily trials of the same individual (Figs 3 and 4). The amount of recovered marker depended on the amount of faeces and the concentration of the marker in the faeces. Both components varied dramatically, although the absence of recovered marker usually reflected the absence of faeces, except during the first hour. In spite of this variability, passage rates showed clear patterns that reflected the amount of intake more than the GI size. Meadow voles had 23% shorter MRTs than did prairie voles, and MRTs for voles fed grass were 21% shorter than for those fed alfalfa (Fig. 5, two-way ANCOVA, P < 0·0001 for species, P < 0·0001 for diet, P = 0·57 for interaction). This pattern is as predicted by intake; greater intake by meadow voles than prairie voles and by both species when fed grass should lead to shorter MRTs. The effect of higher intake must have been greater than the effect of the larger GI volume for voles fed grass, which would otherwise have produced longer MRTs.

Figure 3.

. Examples of temporal patterns for amount of marker (chromium) in faecal collections for two prairie voles fed alfalfa on two consecutive days (a-b for first vole and c-d for second vole). Faecal collections were made at hourly intervals after ingestion of marked fibre. Numbers give estimates of mean retention times. Arrows indicate occurrence of minima in concentration of chromium in faeces. Gaps indicate no faeces or no marker (gaps with arrow).

Figure 4.

. Examples of temporal patterns for amounts of marker (chromium) in faecal collections of two meadow voles fed grass on two consecutive days (a-b for first vole and c-d for second vole). Faecal collections were made at hourly intervals after ingestion of marked fibre. Numbers give estimates of mean retention times. Arrows indicate occurrence of minima in concentration of chromium in faeces. Gaps indicate no faeces or no marker (gaps with arrow).

Figure 5.

. Mean retention time of undigested fibre by prairie voles (MO) and meadow voles (MP) fed grass (hatched) or alfalfa (open). Bars show means and vertical lines show 95% confidence limits.

Discussion

In general, the results supported the IPR hypothesis. As predicted by the hypothesis (see Introduction), both species of voles responded to the more fibrous, less digestible diet (grass) by increasing intake, and both species maintained a constant intake of digestible dry matter on both diets (Fig. 1). Furthermore, both species showed increased GI surface area and mass per unit length (Fig. 2), which indicated an increased radial distention of the GI tract (there was no increase in length) and increased epithelial mass (increased absorptive capability) when fed grass. Both species also responded to the grass diet with faster passage rates (shorter MRT, Fig. 5), which indicated that increased GI size did not entirely compensate for increased intake, and without greater absorptive capabilities neither species could have maintained their required digestible dry matter intake on grass. To the authors’ knowledge, no previous investigation has used natural diets and measured intake, digestibility, GI size and passage rates on the same animals. For this reason, it is believed that this experiment is the strongest test of the IPR to date, and it suggests that all aspects of the IPR are important for maintaining a constant digestible nutrient (or energy) intake on diets with different fibre content.

The responses of surface area and volume to diet were similar, except for the small overall response in the total GI volume of meadow voles to diet (Fig. 2b). The small response of meadow voles to grass could simply reflect errors in calculation caused by the assumption that all compartments were cylindrical. However, the GI contents of meadow voles did not increase either (Fig. 1d). Given that meadow voles showed a 24% increase in daily intake on grasses and a 21% decrease in MRT, the average GI contents and volume probably did not need to change much. Prairie voles, on the other hand, showed a 23% increase in daily intake on grass, but only a 15% decrease in MRT, which suggests that the GI contents and volume should both increase, which they did.

The two species had some additional differences in their nutritional characteristics and responses. Meadow voles had a substantially greater daily intake of digestible dry matter than did prairie voles, which was consistent with previous results and the observation that meadow voles are more active and presumably have higher metabolic rates (Batzli et al. 1994). Meadow voles take more monocots in their natural diet than do prairie voles (Haken & Batzli 1996), and the meadow voles digested grass better, whereas the prairie voles digested alfalfa better (Fig. 1b). Apparently, the ability of each species to digest a food type reflected their relative use of that type in the field.

Differences in the GI tracts of the two species may explain their different abilities to digest grass and alfalfa. Meadow voles had a greater length of the small intestine and a greater total length of the GI tract on both diets compared to the prairie voles (Table 1, Fig. 2a). The greater length of the small intestine may increase the absorptive capacity and may explain the meadow vole's greater ability to digest grass. The high intake required by meadow voles resulted in more rapid passage rates (Fig. 5) because their GI volume was similar to prairie voles, so increased absorptive capacity assumes an even greater significance for meadow voles.

The reliance on increased absorptive capacity rather than on increased GI volume and the reduced retention time of fibre for the meadow voles fed grass suggests that caecal digestion of fibre does little to improve energy acquisition. Arvicoline rodents reingest their faeces, particularly those formed from caecal digesta. Prevention of coprophagy in voles results in only a slight reduction in digestibility of their diet and no change in intake of digestible energy (Cranford & Johnson 1989). Furthermore, reingestion of faeces reduces the amount of fresh food that can be processed by the GI tract. Therefore, it seems more likely that caecal digestion of fibre and coprophagy, which appear necessary to maintain the body mass of voles on low quality diets, serve more to provide essential nutrients than to provide energy (Barnes 1962; Hornicke & Björnhag 1979; Cranford & Johnson 1989).

Prairie voles responded to the grass diet with a substantial increase in GI size and contents, which probably enabled them to utilize grass, even though they prefer dicots. The selection of leaves in preference to the more fibrous stems of plants and the longer MRTs may explain the greater digestion of alfalfa by prairie voles than by meadow voles.

Diet had no significant effect on the total length of the GI tract of either species of vole in this experiment, though caecal length did increase (10% for prairie voles and 25% for meadow voles) with a more fibrous diet (Fig. 2c, Table 1). Total GI dry mass responded to the higher dietary fibre (12 and 45% increase in total dry mass in meadow voles and prairie voles, respectively; 29 and 35% increase in dry mass per unit length of GI tract for meadow voles and prairie voles, respectively), but stomach mass did not (Fig. 2d, Table 1). Gross et al. (1985) and Hammond & Wunder (1991) reported similar results for prairie voles (low fibre diets 7·8% and 8·5% ADF and high fibre diets 39·8% and 22·5% ADF, respectively), although they did not calculate dry mass per unit length. Other studies have reported substantial increases in intestinal length in response to greater fibre in the diet. Hansson (1985) noted that bank voles, Clethrionomys glareolus Schreber, collected from the field where they ate a mixed diet, had total intestinal lengths 10–30% greater than those of voles raised in the laboratory for several months on mouse pellets (largely composed of seeds, a low fibre diet probably about 8% ADF). Later, Hansson & Jaarola (1989) reported similar results for the field vole, Microtus agrestis (L.), which had 10–25% longer intestinal length when collected from areas where it ate a higher fibre diet. The largest differences in intestinal length and mass reported thus far were a 48% increase in length and a 45% increase in dry mass for bank voles fed French bean leaves (13·4% ADF) compared to bank voles fed oat seeds (2·7% ADF) for 28 days (Lee & Houston 1993b). However, under the same conditions, the same authors reported only a 20% difference in total GI length with no significant difference in dry mass for the field voles and no significant differences in either GI length or mass for the much larger water vole, Arvicola terrestris L. Taken together, these results suggest that the strengths of responses of GI length and mass to fibre content of the diet depend on the species and size of vole tested, the levels of fibre in the test diets and the length of time that animals are maintained on the test diets.

A comparison of the few studies that have estimated mean retention times of fibre in arvicoline rodents also shows substantial differences. The earliest study used fuschin to mark epidermal cells of green forage (a mixture of high fibre monocots and low fibre dicots), fed it to three species of voles, Microtus arvalis (Pallas), M. agrestis (L.) and Chionomys (= Microtus) nivalis (Martins), at room temperature, and reported MRTs of 2·2–3·2 h (Kostelecka-Myrcha & Myrcha 1964). Hammond (1989) used mordanted Cr to mark low fibre (8·5% ADF) and high fibre (22·5% ADF) laboratory diets of prairie voles (mean body mass 46–56 g) and found much longer MRTs of 3·8–9·5 h. The lowest values were for voles fed the high fibre diet in the cold (high intake) and the highest values were for voles fed the low fibre diet at room temperature (low intake). From experiments using low fibre (7·4% ADF) laboratory chow marked with mordanted Cr, Hume, Morgan & Kenagy (1993) estimated an MRT of 13·1 h for Townsend voles, M. townsendii (Bachman), at room temperature, much higher than Hammond's value (9·5 h) for comparable conditions. The estimates of MRTs from the present study, all measured at room temperature, ranged between 6·0 and 9·1 h with the lowest values for meadow voles on the grass diet (31·4% ADF) and the highest for prairie voles on the alfalfa diet (19·8% ADF). The results for prairie voles were similar to Hammond's for the low fibre diet at room temperature (9·1 and 9·5 h, respectively) but not for the high fibre diet at room temperature (7·7 and 4·0 h, respectively).

Several factors could explain these differences in mean retention times. First, smaller animals should have shorter MRTs. However, Hammond (1989) and Hume et al. (1993) used voles of similar size (47–56 g and 55 g live body mass, respectively) and got different results. Kostelecka-Myrcha & Myrcha (1964) did not indicate the mass of their voles; they may have been somewhat smaller (30–40 g), although not small enough to account for such short MRTs. The voles used in this study were smaller than Hammond’s, but longer MRTs were found for the high fibre diet. Second, MRT should decrease as digestibility of the diet increases because intake will be lower. However, Hammond and Hume et al. reported similar dry matter digestibilities for their low fibre diets (76·8% and 77·4%, respectively), even though the MRTs were different, and the low fibre diet used was slightly less digestible, but had slower passage than Hammond's diet. Third, passage rates may be faster for larger particles. Björnhag (1994) reviewed the evidence that bacteria and small food particles are selectively returned from the colon to the caecum in various herbivorous mammals, including arvicoline rodents. Although Hammond marked smaller particles (< 0·5 mm) than Hume et al. (0·5–1·0 mm), voles in the genus Microtus grind their food finely so that 75–85% of particles by weight pass into their GI tract at ≤ 0·5 mm, whatever their original size (Lee & Houston 1993c). Therefore, the difference in passage rates seems unlikely to be explained by differences in particle size. Fourth, MRT should become shorter with increased intake. However, Hume et al. reported greater intake than did Hammond for low fibre diets at room temperature (8·1 g and 5·9 g dry matter per day, respectively) and still reported a longer MRT. Fifth, reingestion increases MRT for fibre because it extends the period over which the marker is recovered (Piekarz 1963), and irregular faecal reingestion and deposition increase the variability in the pattern of excretion in faeces. Coprophagous rodents ingest faeces directly from the anus, and coprophagy often alternates with deposition of faeces (Kenagy & Hoyt 1980), but different individuals reingest and deposit faeces at different intervals (Figs 3 and 4). Thus, MRT represents the retention time of undigested fibre assuming a smooth function, but faecal output is sporadic and highly variable, which no doubt leads to some error. The amount of reingestion also increases as the quality of the diet decreases (Kenagy & Hoyt 1980; Cranford & Johnson 1989; Björnhag 1994).

Finally, MRT based on indigestible fibre should not be viewed as representative of the diet as a whole, most of which is digested and absorbed or defaecated during the first passage. If fuschin is lost from reingested fibre but chromium is not, the shorter MRTs reported by Kostelecka-Myrcha & Myrcha (1964) for dyed particles would be expected because their MRTs (2–3 h) would represent a single passage and would be more similar to turnover time (mass of GI contents/rate of intake). Turnover times calculated from the data in the present study yielded low values of 0·7–1·6 h. Thus, the discrepancies among different studies using a chromium marker could result from different amounts of faecal reingestion in different experiments. Reingestion of faeces could also explain the otherwise puzzling result of Hammond & Wunder (1991) that voles fed high fibre diets showed decreased overall dry matter digestibility but increased the digestibility of fibre itself. It is concluded that future work on passage rates in voles should be based on standard methods that take into account the reingestion of faeces so that valid comparisons of MRT can be made among different species and studies. In fact, measurement of a single passage or of turnover time might be preferred to mean retention time of undigested fibre when the objective is to characterize the rate of processing of the diet as a whole.

Acknowledgements

We are grateful to A. Kosowski, R. Perlot, C. Lacelles, N. Siddiqui, and C. Javier for their help in conducting the laboratory research. Dr M. David graciously allowed us the use of his atomic absorption spectrophotometer. S. Smothers helped to ensure the accuracy of our analyses. Financial support for this work was provided by a grant from the University of Illinois Research Board and by a University of Illinois Chancellor Postdoctoral Fellowship to the senior author.

Footnotes

  1. Present address: Departments of Anthropology and Biological Sciences, California State University, Long Beach, CA 90840, USA

  2. For correspondence.

Ancillary