Serial sampling provides chronological evidence that endogenous protein is used for primary growth in a molt-migrant goose

This is a proof of concept paper based on chronological samples of growing feathers from geese thought to be molt-migrants. When molt-migrant birds initiate molt shortly after migrating to a new isoscape, isotope values measured along the length of their feathers should change continuously. To assess long-term changes and daily cycling in δ15N and δ13C values, we serially sampled growing primaries of three presumed molt-migrant geese. Two showed changing δ15N signatures along the length of their growing primaries, indicating they were molt-migrants, while the third, presumably a resident, showed no change. We then resampled these feathers at closer intervals for evidence of the predicted diel cycle in the use of exogenous and endogenous protein for feather growth, generated by the diel feeding cycle of these geese. As predicted, the two geese that were equilibrating to a new isoscape showed oscillations of approximately 24-hour periodicity in δ15N values, measured along the length of their primaries. In contrast, the goose that was not equilibrating to a new isoscape showed no 24-hour periodicity in its δ15N values. Our results demonstrate that chronological sampling along the length of individual primaries holds great potential for identifying individuals that are molt-migrants.


INTRODUCTION
Correspondingly, stable isotope values in growing flight feathers of Greylag Geese were intermediate between birds on a terrestrial plant diet and those on a saltmarsh diet (Fox et al. 2009).
To explore expected changes in δ 15 N and δ 13 C along the length of primaries and to examine the potential for 24-hour cycling in the use of exogenous and endogenous protein for feather growth, we serially sampled a single primary feathers from each of three Greylag Geese from Saltholm; these feathers (and no others) were available from the study of Fox et al. Our results show that serial samples taken along the length of primary flight feathers readily show equilibration toward a new isoscape and, further, that feathers sampled at fine intervals can reveal diel shifts in exogenous and endogenous protein sources used for feather generation. These results clearly demonstrate that the chronological samples available from individual feathers can be used to identify individuals that are molt migrants. We are aware of just two prior studies of hair or feathers that used chronological samples to examine changes in isotope signatures through time. Cerling et al. (2006) used elephant hair to demonstrate the movement of individual by repeated rinsing. After air drying the feathers were cleaned again in 2:1 chloroform:methanol and allowed to air dry before processing. Once feathers were dried a research technician marked the rachis of each feather from its tip to the base of the growing vane at 1, 2 or 5mm intervals. The majority of the posterior vane was then cut away, leaving only the few mm closest to the rachis. Then, using the pen marks as a guide, an approximately 200μg sample of feather vane was cut immediately adjacent to the pen mark.
These samples were loaded into tin capsules (3.5x5.5mm) and stored in an elisa plate until they were analyzed. Among sample variation in the laboratory reference material was < 0.2‰ for both δ 13 C and δ 15 N. In addition, to this laboratory reference we ran one sample each of two National Institute of Standards and Technologies NIST reference materials (USGS 40 in autorun position 25 and USGS 41 in autorun position 31). All stable isotope ratios are expressed in standard δ notation, where δ 13 C and δ 15 N = [(isotope ratio sample/isotope ratio standard) -1] * 1000. Consequently, δ 13 C and δ 15 N are expressed in parts per thousand (‰) deviations from a standard, which was Vienna Pee Dee Belemnite for δ 13 C and air for δ 15 N. Isotope ratios were measured at the University of Oklahoma using a Thermo Finnigan Delta V isotope ratio mass spectrometer connected to a CosTech elemental analyzer. To determine if there was evidence of 24-hour cycling in the δ 15 N values along the length of the primaries, we used custom Matlab code to perform an autocorrelation analysis after de-trending the data using linear regression and removing the mean. We used this method to find correlation maxima and minima that reveal periodicity in the δ 15 N values. Using additional custom Matlab code, we then developed a bootstrap method to test for the statistical significance of having autocorrelation minima and maxima that correspond to a periodic pattern in isotope values. To do so, we randomly permuted the data for each feather and performed autocorrelation analyses of those permuted values. Out of 10,000 permutations, we asked what fraction of the data had both a minimum less than or equal that observed in our original autocorrelation and a maximum spaced at the appropriate interval.

Evidence for equilibration following molt-migration
δ 15 N signatures declined over time in two geese (501 and 508, P < 0.0001), while 509, showed no change (P = 0.34, Fig. 1). Both geese that changed did so in a way consistent with the large shift of about 8 ‰ in the δ 15 N isoscapes suggested by the results of Fox et al. (2009). δ 15 N was constant, with a mean of 7.6 ‰, along the primary for goose 509, which is puzzling because this mean is intermediate between the beginning and ending values for the other two geese (Fig. 1). If this individual had been on Saltholm long enough to be in equilibrium with the salt marsh isoscape, then its mean δ 15 N value should have been at or below the latest values from the two geese that showed declining δ 15 N values. That its mean δ 15 N was considerably higher than the lowest δ 15 N values found for the two geese coming into equilibrium (Fig. 1), suggest it was a resident goose that did not feed in the Saltholm salt In general, measuring isotopic changes in chronological samples taken at equal intervals from flight feathers offers a powerful tool for studying moltmigration. It provides strong data for individual birds while avoiding the assumptions involved with food sampling and using fraction estimates to compute expected tissue values for isotopes. Serial samples representing equal time intervals through primary growth can be generated in two ways.
For large birds, finely spaced samples along the length of a primary are chronologically so accurate that they can be used, not only to assess isotopic change during feather growth, but also to evaluate the expected 24-hour periodicity in isotope measurements driven by foraging schedules. For small birds with short primaries, serial samples from single primaries would generate only a limited temporal series and their primaries grow so slowly (Rohwer et al. 2009) that sampling with laser ablation would be required to achieve a sample density sufficient to detect 24-hour cycling (Moran et al.