Skip to main content
Dryad logo

Advance social information allows red crossbills (Loxia curvirostra) to better conserve body mass and intestinal mass during food stress


Cornelius, Jamie (2022), Advance social information allows red crossbills (Loxia curvirostra) to better conserve body mass and intestinal mass during food stress, Dryad, Dataset,


Animals prepare for fluctuations in resources through advance storage of energy, planned reduction in energy costs or by moving elsewhere. Unpredictable fluctuations in food, however, may be particularly challenging if animals cannot avoid negative impacts on body condition. Social information may help animals to cope with unpredictable resources if cues from individuals with low foraging success give advance warning about deteriorating conditions. This study investigates the impact of social information on behavior and physiology of food-restricted captive red crossbills (Loxia curvirostra). Birds were restricted to two short feeding periods per day to simulate a decline in resources and were given social information from food-restricted neighbors either before (i.e., predictive) or during (i.e., parallel) the food restriction period. Focal birds better conserved body mass during food restriction if social information was predictive of the decline in resources. Crossbills with predictive information ate more food, had larger intestinal mass and better conserved pectoral muscle size at the end of the restriction period compared to those with parallel social information. These data suggest that birds can use social information to alter behavioral and physiological responses during food shortage in ways that may confer an adaptive advantage for survival.


This experiment was organized into a preparative phase during which focal birds received social information from well-fed or restricted neighbors, followed by a restrictive phase during which focal birds were food restricted. Focal birds were provided with social information from food-restricted neighbors for either three days prior to their food restriction (i.e., Social Predictive focal group) or for three days coincident with the focal food restriction (i.e., Social Parallel focal group). The total amount of social information received about declining food was thus similar between the two focal groups; however, this design also resulted in different social information contexts during the restrictive phase: predictive focal birds had well-fed neighbors whereas parallel focal birds had restricted neighbors. To control for this, I also compare the response to food restriction in the neighbor groups. Neither of the neighbor groups received predictive information prior to their respective restrictions; however, the Predictive Neighbors were housed next to well-fed birds during their restriction whereas the Parallel Neighbors were housed next to restricted birds.

Capture and captive conditions

The red crossbill species is a complex of different eco-types that are morphologically adapted to particular species of conifer seeds and which can be acoustically identified by type-specific vocalizations. This experiment utilizes Type 2 red crossbills, which are found frequently in ponderosa pine (Pinus ponderosa) dominant habitats. Ninety-six Type 2 red crossbills were captured via mist net at various field sites in the east Cascade mountain range of Oregon and Washington, USA in late summer of 2019. Birds were held in captive facilities at Oregon State University on natural photoperiod and room temperature, and with ad libitum pellet food and daily sunflower seed treats until the experiment started the following summer. All birds were handled once per week for a separate experiment that monitored seasonal changes in body condition from March to early June. A subset of birds (N = 24) experienced a long-day photoperiod exposure for several weeks in late March (see supplement of publication for details). These birds were split equally between the Predictive Neighbor and Parallel Neighbor groups and we detected no differences in initial mass or in their response to food restriction compared to individuals that had not previously been photoadvanced.

Social Information

Sex and age-at-capture were balanced across all treatment groups in three experimental rooms. Focal birds were housed alone in a small cage (42 cm L x 30 cm W x 56 cm H) with visual contact with one pair of neighbors housed in a cage next to them. Neighbors were the same sex as the focal bird and included one juvenile and one adult in each pair. Control individuals were visually separated from other birds but could hear the other control birds. Similarly, each social group could hear the other individuals of their treatment group only (i.e., one treatment group per experimental room). 

Food Restriction and Food Intake

Birds were fed a commercial maintenance pellet food designed for songbirds (see supplement for nutritional information). Controls had access to food ad libitum for the entire experiment. Focal and neighbor birds in the social groups each experienced 3 days of food restriction during which they had access to food for two 45-min feeding sessions per day. Food was provided in a covered cup to minimize spills and food intake was calculated as the change in grams of food remaining in the cup across a 24-hour period, measured with an electronic balance to the nearest 0.01 g. Multiple food cups were provided in neighbor cages to ensure both birds had access to food during the feeding periods. One bird in the Social Predictive group regularly flung food out of its cage, thus food intake was not calculated for this bird.

Body Condition Metrics

Body mass was measured with an electronic balance to the nearest 0.01 g. Muscle and fat deposit scores were collected by a single experienced observer to the nearest 0.25 scale unit in a partially blind design. The observer knew the stage of the experiment and from which room data were being collected, but did not know the bird’s individual identification or if the bird was from a neighbor or focal group at the time of scoring. Pectoral muscle size was scored on a scale of 1 to 4 based on the shape of the muscle around the keel, described by Bairlein in detail. Fat deposits were scored in the abdominal and furcular cavities based on Helms and Drury.

Organ Mass and Carcass Lipid Composition

Six randomly chosen individuals from the Control, Social Predictive and Social Parallel groups were euthanized by overdose of isoflurane, plucked free of feathers and immediately dissected for wet organ mass measurements immediately following the Day 3 sampling session. Another six individuals from the Social Parallel group were sacrificed later after having been refed on ad libitum food for two weeks to determine how intestinal mass in food restricted birds responded to refeeding and to help interpret the observed patterns during restriction. I did not sample from all groups to reduce the number of birds sacrificed. Wet masses of liver, gizzard, esophagus, stomach, intestine and colon were collected using an electronic balance to the nearest hundredth of a gram. Alimentary organs were each cleared of contents by cutting each structure open with surgical scissors, rinsing with water and then blotting dry prior to measurement. Organs were immediately placed back in the carcass and the carcass was frozen at -80° C . Percent lipid in dried carcass mass was determined by Soxhlet extraction (see supplement for details) to test if visual scoring of subcutaneous fat stores predicted whole-body lipid composition.


Activity was monitored continuously throughout the experiment with infrared activity monitors from STARR Life Sciences (Oakmont, PA, USA). These monitors record the number of beam breaks over time and capture whole body movements using an array of infrared beams projected throughout the cage. Sensor ID and position on the cage were kept constant for each individual throughout the course of the experiment to minimize any impact of detection variability on changes in activity levels. Monitor position ensured that activity was captured from only a single focal cage.  Three sensors failed during the experiment (two in the Social Predictive and one in the Social Parallel). 


National Science Foundation, Award: 1755227