Wild birds must consume certain amounts of protein and an appropriate balance of amino acids while inhabiting environments where foods often differ in the quantity and quality of available protein. The requirements for amino acids are well documented for domestic bird species but are largely unknown for wild birds, which makes it impossible to reliably assess the nutritional adequacy of foods eaten by wild birds. We measured the maintenance requirements for three essential amino acids (lysine, methionine, and arginine) in two species of songbird, the omnivorous Hermit Thrush (Catharus guttatus) and granivorous White-throated Sparrow (Zonotrichia albicollis). Hermit Thrushes and White-throated Sparrows had similar requirements for lysine (20.02 and 19.95 mg/day, respectively) and methionine (12.3 and 10.85 mg/day, respectively), whereas thrushes had lower requirements for arginine (18.07 mg/day) compared to sparrows (34.5 mg/day). Consistent with previous studies, most birds fed diets with inadequate essential amino acid concentrations reduced food intake and fecal output, lost body mass, and had lower, but not negative nitrogen balance. However, we provide the first evidence that songbirds overcompensate when they consume diets very deficient in lysine. Available data on amino acid concentrations in natural foods suggests that most insects contain relatively high concentrations of all essential amino acids, seeds likely satisfy requirements of lysine and arginine but not methionine for Hermit Thrushes and White-throated Sparrows, whereas fruits generally contain inadequate amounts of all essential amino acids. Therefore, birds that eat mostly fruit may consume enough protein but likely must eat other types of foods to satisfy their essential amino acid requirements.

Capture and maintenance of birds

We used mistnets to capture white-throated sparrows (n = 21) and hermit thrushes (n = 20) during fall migration in 2006 in Kingston, Rhode Island (41°28’N, 71°31’W). The birds were transferred to indoor facilities and housed individually in stainless-steel cages (59´45´36 cm) at constant temperature (23°C) and a constant photoperiod representative of natural photoperiod at capture (12 h light:12 h dark, lights on at 0800 hours). Following capture, a 10-week acclimation period provided birds with ad libitum food and water along with about 8 to 10 mealworms (Tenebrio molitor) each day. All experiments were run within six months of the initial capture date.

The nutrient composition of the acclimation diet was similar to that of many natural high-carbohydrate fruits (59.5% carbohydrates, 12.8% protein (casein) and 8.0% fats; see Table 1 Langlois and McWilliams 2010). Songbirds fed similar semi-synthetic diets have been successfully maintained for more than one year (Murphy and King 1982; McWilliams et al. 2002; Pierce and McWilliams 2004; Pierce and McWilliams 2005; Casagrande et al. 2020). The acclimation diet was formulated so that the essential amino acid concentrations satisfy the maintenance requirements of white-crowned sparrows (Online Resources Table 1 and 2) (Murphy 1993b). Body mass was measured daily (± 0.1 g) and birds remained in good health. 

Diets and experimental design

After the 10-week acclimation period, we offered birds only the casein-based acclimation diet (no mealworms) for the next 6 weeks. Birds were then used in a protein requirement experiment (Langlois and McWilliams 2010) for nine days. Thereafter, birds were switched to a crystalline amino acid-based diet (diet A for all EAA experiments; Online Resources Table 1). Diet A was nearly identical in composition to the acclimation diet except L-crystalline amino acids replaced casein as the sole source of dietary protein. This allowed for the adjustment of minute quantities of individual EAAs to create diets with varying amounts of EAAs, which was not possible with the casein-based diet. 

We randomly assigned birds to one of four initial diet groups (A, B, C, D) that differed in EAA concentration. Therefore, birds were fed one of the four diets during the first EAA trial but rotated to a different dietary level (A-D) in subsequent trials to reduce potential carryover effects. Diets were isocaloric and nearly isonitrogenous because we replaced the EAA (lysine, methionine, or arginine) with reciprocal amounts of glutamic acid (Online Resources Table 2). The dietary EAA concentrations were chosen so that birds in some groups (diets A and B) were fed diets intended to provide adequate EAA whereas birds in other groups (diets C and D) were fed diets intended to provide inadequate EAA given the EAA requirements of white-crowned sparrows (Murphy 1993b; Fig. 1). We used Murphy (1993b) as a guide to determine the exact EAA concentrations in our experimental diets.

We conducted 3-day total-collection trials (Murphy 1993b) for each diet group of each EAA trial for both species during the 18^th through 21^st week of captivity in the following order: lysine, methionine, arginine, and lysine repeated (lysine-repeat). Each trial was separated by 3-4 days on the maintenance diet (diet A). Two identical lysine trials (lysine and lysine-repeat) were conducted with each bird to confirm that there were no changes in the bird’s estimated EAA requirements over time. For the first trial (lysine), we randomly assigned 4-6 birds of each species to one of the four diet groups (A-D) that differed in EAA concentration. To reduce potential carryover effects between subsequent trials, we reciprocally switched birds fed with the highest EAA amount in the first trial (diet A) to the diet (D) with lowest EAA amount in the next trial and visa versa. We also reciprocally switched birds fed diet B in the first trial to diet C in the second trial and vice versa. For the third and fourth trials, birds returned to their original randomly selected diet group (A-D). Two hermit thrushes were removed from the experiment after the first lysine trial because they were behaving abnormally, although there were still 4-6 birds in each diet group. At 0800 hours each day during these 3-day trials we measured each bird’s body mass, provided each bird with ad libitum fresh food and water, and weighed the food that remained from the previous day. We also collected all excreta produced by each bird during the previous 24 h, along with samples of food offered and remaining. All samples were stored frozen at -20 °C for later analysis.

We dried food remaining at 47 °C until sample mass was constant (~1 week). The samples of food offered and collected excreta that were used for measuring nitrogen content were freeze-dried until constant mass (2 days) to ensure that nitrogen was not lost during drying. Dried samples were homogenized using mortar and pestle. Methods for measuring nitrogen concentration of food and excreta are described in Langlois and McWilliams (2010).

We estimated nitrogen balance (mg day^-1) as the N intake minus N lost in excreta; the equation is as follows: N Balance = (F ´ FN) – (D ´ DN) where F is amount of food consumed (g DM  day^-1), FN is nitrogen content of the food (mg N/ g DM), D is the amount of excreta (g DM day^-1), DN is the nitrogen content of the excreta (mg N/ g DM) (Murphy 1993a). Because all EAA diets were nearly isonitrogenous, we did not expect energy density in food or excreta to change during the trials and we confirmed this by measuring energy density of excreta in the lysine trials (Langlois 2008).

Statistical analysis

We conducted three analyses to confirm key aspects of our experimental design (see Online Resources for details): 1) adequate acclimation time within the 3-day collection trials, 2) no change over time between the two lysine trials conducted at the start and end of the experiment, and 3) confirming that diet A was nutritionally adequate. After confirming these key aspects of the experimental design (results in Online Resources), we used ANOVA for hermit thrushes and white-throated sparrows fed diets A - D to determine the effect of EAA intake on N balance, food intake, amount of excreta, and percent nitrogen in excreta for each species on day 3 of each EAA trial. We used a conservative P value (0.01) for Levene’s Test of Homogeneity of Variance due to the robustness of the ANOVA model; all reported statistical analyses satisfied the assumptions of normality and homogeneity of variance among treatment groups. 

We used linear regression to estimate birds’ requirements for each of the three EAAs. For each species and EAA, we regressed change in body mass and N balance on EAA intake, and food intake on dietary EAA concentration for birds fed deficient diets on day 3 of each trial. Diets were considered deficient when birds had significantly lower N balance than the grand average N balance for birds fed adequate EAA (see Figs. 2 thru 4), or when birds lost significant body mass. We used t-tests to compare slope and elevation (y-intercept) parameters from these regression equations between hermit thrushes and white-throated sparrows in all EAA trials (Sokal and Rohlf 1981). Values are reported as means ± SD and the significance level was set at P £ 0.05. All statistical analyses were conducted using R version 4.0.0 (R Core Team 2020).