Hydration capacity: A new criterion for feed formulation

• We measure hydration capacity of 24 raw materials and 28 compound pelleted feeds.
• Hydration capacity is highly variable between the 24 raw materials tested.
• Grinding process influence hydration capacity.
• Hydration capacity of a compound feed increases with those of its ingredients.
• Predictability of hydration capacity of a feed from its ingredients is moderate.

Twenty-four raw materials, harvested or produced in 2013 and commonly used in animal feed, were used to measure their hydration capacities through both water-holding capacity (WHC, g H2O/g DM after 24 h of water addition) and swelling capacity (SC, mL H2O/g DM during 60 min after water addition, every 5 min during the first 30 min and then every 10 min). The raw materials were provided in the form of whole seeds (F1 class; n = 8), flour/mash/hulls (F2 class, n = 8) or pellets (F3 class, n = 9). Hydration capacities were measured for unprocessed material and also after grinding of whole seeds (1, 3, 5 or 8 mm grid). The raw materials were analysed for fibre content to study the relation with WHC and SC. Moreover, WHC and SC were measured on twenty-eight compound pelleted feeds containing some of the raw materials studied in the present experiment, harvested or produced in the same year (i.e. 2013; Experimental feeds; n = 8) or in an earlier year (i.e. <2013; Reference feeds; n = 20) to evaluate additivity and predictability of these criteria. WHC and SC at 60 min (T60) varied greatly among the raw materials (2.88 ± 1.74 g H2O/g DM and 2.61 ± 2.37 mL H2O/g DM, respectively; P < 0.001) and are weakly correlated (R2 = 0.52; P < 0.001). They are thus two complementary measurements interesting in feed formulation. The physical form of the raw materials at the moment of delivery had a significant effect on WHC (0.97, 4.92 and 3.65 g H2O/g DM in the F1, F2 and F3 classes, respectively; P < 0.001). Grinding the seeds had a significant effect on WHC and SC values (P < 0.001). WHC and SC at T60 were moderately correlated with the fibre content: neutral detergent fibre (aNDF), acid detergent fibre (ADF), acid detergent lignin (ADL), hemicellulose and cellulose content (R2 < 0.40 and P < 0.01 the same for every fibre fraction). In the eight Experimental feeds, the correlation between calculated SC and measured SC at T60 was low (R2 = 0.38; P = 0.104). Conversely, the correlation between calculated WHC and measured WHC was high (R2 = 0.89 and R2 = 0.81; in the Experimental and the Reference feeds, respectively; P < 0.001) and the mean difference between the predicted WHC and the measured WHC was moderate (9.7% and 10.6%, respectively for Experimental and Reference feeds), but could reach high values (>15%), especially when measured WHC was high. In conclusion, hydration capacity could be considered as additive, however, the theoretical values of pelleted compound feeds should be supported by real measurements.