For heifers fed a backgrounding diet, fecal DM concentration (%) differed over 24 h (P = 0.01) with DM being the lowest for the 0- to 4-h and 20- to 24-h intervals relative to feeding.
Fecal NDF and ADF also showed variation over 24 h, with greatest output occurring between 12 to 16 h. Fecal constituent concentrations were not affected by feeding frequency (P > 0.05), and there were no interactions between feeding frequency and time (P > 0.05). Measured fecal DM, NDF and ADF concentrations at specific individual sampling times differed (P > 0.05) from those calculated using total collection, whereas no differences were observed for the other constituents (Table 4.4).
For heifers fed the finishing diet, greater 24-h variation in fecal constituents was observed, and except for fecal N, all fecal constituent concentrations measured at individual sampling times displayed some differences (P > 0.05) from those derived using total collection (Table 4.4). Fecal DM, starch, NDF, ADF, and ADL concentrations differed over 24 h (P < 0.05), with a trend (P = 0.05) observed for OM (Table 4.4). Fecal DM (%) was lowest from 0 to 4 h, and highest from 8 to 12 h, and both intervals differed from the average DM concentration determined from total collection. Fecal starch concentration peaked at 4 to 8 h, and was lowest from 16 to 24 h, and differed from the total collection value at these time intervals. Compared to fecal starch, changes in the other fecal constituents were more continuous and did not display as many dramatic increases or decreases. None of the fecal constituent concentrations were affected by feeding frequency, but interactions between feeding frequency and time were observed for fecal N, NDF and ADF. Fecal N, NDF and ADF concentrations were more variable over 24 h in FF1 heifers compared to FF2, and no differences were observed for fecal N excretion in FF2 heifers. When collected from FF1 heifers, fecal NDF and ADF concentrations differed from their total
collection value in the evening (8 to 16 h after feeding for NDF and 12 to 16 h for ADF). Fecal ADL also differed from the total collection value in the evening samples, 12 to 20 h after first feeding.
There has been limited work to characterize the diurnal excretion patterns of dietary constituents in ruminants, with most studies describing variation in the excretion of starch (Leonard et al. 1989; Caetano 2008; Fredin et al. 2014) and digestibility markers such as lignin (Kane et al. 1952; Kanani et al. 2012). Fredin et al. (2014) sampled feces from lactating cows fed alfalfa silage and corn silage in equal proportions ad libitum, and two different types of grain
mixes containing dry ground shelled corn either at 15.4% or 24.4% of the diet DM as described in Akins et al. (2014). Fecal samples were collected at 0 and 12 h after feeding, with greater fecal starch concentrations found in the morning (2%) as compared to the evening (1.6%). Caetano (2008) sampled hourly from four Nellore steers fed ground corn as the sole starch source (28.5%
corn in the diet DM) once daily. A polynomial pattern of starch excretion (fecal starch = 0.12 - 0.0074 × time + 0.00026 × time2) was observed, with the greatest fecal starch levels measured in the morning (mean 7.5%), followed by a drop 10 to 18 h after feeding (mean 2.3%). Leonard et al. (1989) measured changes in fecal starch concentration over three 8-h periods from Angus steers offered whole corn (2.27 kg/d) and chopped fescue hay for ad libitum intake once daily.
The greatest concentrations of fecal starch were also observed in the morning (21.3% in the first 8 h of sampling) as compared to later in the day (mean of 13.4% in the remaining 16 h). In the same study, grinding corn resulted in more fecal starch being excreted in the evening (10.3% at 16 to 24 h) compared to the morning (7.76% at 0 to 8 h, and 4.83% at 8 to 16 h). These findings suggest that the degree of grain processing affects the optimal fecal sampling time for
estimations of starch concentration or digestibility.
In the backgrounding period, fecal starch concentration did not vary significantly over 24 h.
Although the level of grain fed in the backgrounding period (44% of dietary DM) was greater than in previous studies with corn (Leonard et al. 1989; Caetano. 2008; Fredin et al. 2014), barley is more digestible than corn, and could explain low fecal starch levels and the absence of diurnal fluctuations. Another possibility for the failure to observe statistical differences could be a result of large residuals of analysis. For low starch values, the average C.V. for duplicate samples were much higher compared to wet chemistry (0.33 versus 0.055), indicating that more replicates should be used when predicting samples where fecal starch is expected to be low (≤
2% fecal starch). When fecal starch values increased (as observed in the finishing period), the average C.V. for NIRS analysis were only slightly higher than wet chemistry (0.067 versus 0.053). Due to the possibility for variability, closer inspection of statistical differences for feeding frequency and time effects was done. All differences between feeding frequencies and between time of sampling fell well within the 95% confidence intervals, however the PDIFF estimate between 0 to 4 h and 4 to 8 h showed a significant difference (P = 0.02). This
inconsistency can occur with small datasets, and although the overall likelihood ratio tests should be more robust than the pairwise comparison tests (Wald tests) (C. Waldner, personal
communication, ), it suggests that with less variability in starch measurements, the difference between these two time intervals may be regarded as significant.
In the finishing period, the increase in starch excreted early in the day aligns with the findings of other studies where greater starch concentration in cattle feces was observed in the morning as compared to the evening (Fredin at al. 2014; Caetano. 2008; Leonard et al. 1989).
Since grain has a smaller particle size and higher specific gravity as compared to forage particles, it has a lower ruminal retention time, potentially leading to more variable excretion over 24 h.
Upon entering the rumen, grain tends to sink and remain in the same location, giving it an early entrance to the reticulum and omasum compared to forage (Schalk and Amadon 1928). Passage of forage is more dependent on contractions of the reticulum and rumen, along with regurgitation and mastication (Schalk and Amadon 1928) and consequently the longer retention time results in a more uniform passage rate and fecal excretion pattern. Upon closer examination of the 95%
confidence intervals, it was noted that the difference between the two feeding frequencies was quite large (2.09% greater fecal starch in FF2 heifers), and the confidence intervals were
exceptionally broad (-8.09 to 3.91) indicating the possibility of an effect of feeding frequency or an interaction between feeding frequency and time on fecal starch excretion. This enforces the need for standardizing sampling times based on different diets and management practices.
Fecal N originates from dietary, microbial, and endogenous sources, and would not be solely influenced by the passage rate of feed particles. Endogenous N in feces consists of sloughed epithelial tissue and undigested residues of enzyme secretion, and is assumed to also be constant per unit DMI (Strozinski and Chandler 1972). Microbial N is related to the large intestinal fermentation of DMI (Mason 1969; Mason and Frederickson 1979), which is more variable in cattle fed a high fermentable diet once per day, explaining the variation observed in N excretion in heifers fed a finishing diet once per day.
Although there are no studies to compare fecal fiber concentrations over 24 h, our results show significant changes in fecal fiber that are of low magnitude, and ones that vary depending on feeding frequency in the finishing period. Unlike low starch concentrations, the ability to measure NDF, ADF, and ADL using NIR are more precise with C.V. values all below 0.026.
The greatest excretion of all fiber fractions and ADL are observed between 8 to 20 h after first feeding, coinciding quite closely to the time of reduction in fecal starch concentration.
High forage diets are consumed more slowly than high concentrate diets with more time spent chewing and ruminating (Bailey 1961; Beauchemin 1991). As a result, excretion of fiber would be expected to be more uniform over the day. High grain diets have a reduced ruminal retention time and are more variable in constituent excretion, particularly of starch. It is likely that in the finishing period, declines in ADF concentration in the morning and its increase in evening reflect the dilution of ADF by starch in the feces in the morning and a reduction in fecal starch in the evening. Since fecal fiber concentrations were more variable over 24 h in cattle fed only once versus twice per day, fecal samples collected from cattle fed multiple times per day at any time point will more accurately represent fiber levels in a 24-h fecal composite.
Kanani et al. (2012) fed cattle bermudagrass hay twice daily with a 9-h interval between meals and measured daily excretion of alkaline peroxide lignin at 6-h intervals. Alkaline
peroxide lignin concentrations in fecal samples were similar across sampling times; a result that is consistent with our observations of the excretion of ADL with 4-h intervals in the
backgrounding period. The concentration of this constituent is greater in forage than in grain, and as we observed, mirrors the excretion of forage particles. In a separate study, diurnal variation of lignin excretion was measured in three dairy cows fed at 0430 and 1330 h (Kane et al. 1952).
The diet consisted of (DM basis) 80% alfalfa hay and corn silage and 20% concentrate. The percentage of lignin in feces differed (P ≤ 0.05) between the morning and afternoon, but did not appear to be influenced by the time of feeding. The optimal time for sampling was proposed to be between 8.5 h to 10.5 h after the first feeding, as this period most closely resembled the average lignin concentration over 24 h (Kane et al. 1952). In contrast, we did not detect differences in ADL concentration between individual fecal samples and the 24-h composite when heifers were fed the backgrounding diet, but we did observe differences in the finishing period, with samples collected between 0 and 12 h after feeding being similar to total collection.
Our results indicate that there is little effect of sampling time on the excretion of fecal constituents in heifers fed the backgrounding diet, and a single sample at any time over 24 h would generate estimates similar to those derived from total collection for OM, starch, N, and ADL. Differences in NDF and ADF were small and may require standardization of timing of sampling if very accurate estimations are required. In contrast to fecal starch, fecal fiber levels have not shown to be of much value for feedlot cattle; hence the differences may be negligible at a commercial setting. Since there were no changes in ADL over 24 h, it is likely that any time
point can be used when applying the marker method to predict digestibility. When fecal starch concentration is expected to be low it is recommended to measure each sample multiple times to generate more accurate and precise results. In contrast, when heifers were fed the finishing diet, different sampling times resulted in different fecal constituent concentrations over 24 h. It is recommended that time of sampling be standardized when collecting fecal samples in order to be representative of total collection, and fecal starch concentration in morning sampling of feces (0-4 h) approximates closely to fecal starch concentrations derived from total collection.
4.3.5 Accuracy of Using Fecal Nutrients to Predict Apparent Total Tract Digestion