1. Anthropogenic subsidies to native predators can have cascading effects on sensitive prey populations, but the spatial mechanisms behind these effects are often unknown.

2. We used a stable isotope mixing model to reconstruct spatially naïve assimilated diets of common raven (Corvus corax) chicks and then used regression analysis to investigate landscape patterns in assimilated chick diet, with particular respect to the eggs and chicks of greater sage-grouse (Centrocercus urophasianus).

3. Assimilated raven diets were primarily composed of mammal carrion, followed by anthropogenic food and sage-grouse eggs and chicks.

4. Raven diets showed landscape gradients, whereby raven chicks in nests near active greater sage-grouse breeding leks consumed a higher proportion of sage-grouse eggs, sage-grouse chicks, and insects in their diet and less mammal carrion. A majority of raven nests on anthropogenic nesting structures (78.7%) were within 5 km of the nearest sage-grouse lek. Ravens nesting in high-probability greater sage-grouse nesting habitat consumed more insects and plants and less mammal carrion.

5. In landscapes devoid of natural raven nesting substrates, such as our study area, anthropogenic nesting substrates can ‘anchor’ breeding ravens nearer to greater sage-grouse leks, with concomitant increases in raven predation on greater sage-grouse nests. Curtailment of anthropogenic nesting substrates within 5 km of a sage-grouse lek may have a disproportionately positive impact on sage-grouse populations. More generally, these findings highlight that the spatial arrangement of anthropogenic subsidies can result in indirect interactions between humans and predators with direct implications for predators and prey.

Stable isotope sampling

Feathers are keratinized tissues that preserve the isotopic record indefinitely and can represent data spanning the period for which the feathers were grown (Pearson et al. 2003).  We collected and combined two scapular feathers from 179 raven nestlings (usually two nestlings per nest) via clipping for stable isotope analysis of carbon (δ¹³C) and nitrogen (δ¹⁵N).  Feathers were collected between May 9^th and July 3^rd in 2013 and May 2^nd and July 17^th in 2014.  Only fully-grown or emerged portions of pin feathers were collected (i.e., dry feathers only) and clippings measured approximately 50 mm on average but varied depending on the age of chicks and stage of feather development.  Feathers were placed in individually labeled and sealed paper envelopes and stored at room temperature until shipment to the analytical lab.  The sampling was non-lethal and minimally invasive.  All raven chicks from which feathers were collected were banded with USFWS leg bands and returned to the nest (Federal Bird Banding Permit #23780).

We also collected samples from potential diet items within the study area. These items included hair samples collected from opportunistic road-killed herbivore mammals (n=6; three domestic cows [Bos taurus], one domestic sheep [Ovis aries], one jackrabbit [Lepus townsendii], and one pronghorn [Antilocapra americana]), greater sage-grouse egg shell membranes from already hatched or depredated nests (n=13), sage-grouse feathers from already hatched or depredated nests (n=5), and sage-grouse feathers from a depredated hen (n=1) found within the study area.  We also collected anonymous human hair samples that were pooled from floors of hair salons and barber shops in Rawlins, Wyoming.  These human hair samples were used to generally reflect composite anthropogenic diets (e.g., composite isotopic signatures of anthropogenic food sources such as dumps and transfer stations often based on C₄ plants; O’Connell and Hedges 1999).  All samples were stored in individually labeled paper envelops and sent to the analytical lab for analysis.      

Laboratory analysis

All samples were sent to the University of Wyoming Stable Isotope Facility (Laramie, Wyoming, USA) for sample preparation and analysis following each field season.  To clean the samples, a 2:1 mixture of chloroform:methanol was used to rinse each sample, followed by 3 rinses of deionized water.  Samples were loaded into tins with a range from 0.75 to 0.85mg.  The standard uncertainty for the lab’s instruments for carbon is 0.15 permil and for nitrogen is 0.2 permil, although the precision was less than or equal to 0.1 permil for all runs of our samples.  Samples were ground into extremely fine powder using a ball mill.  Isotopic analyses were conducted using a Carlo Erba 1110 or Costech 4010 Elemental Analyzer coupled to a Finnigan Delta+XP continuous flow inlet isotope ratio mass spectrometer (Thermo Finnigan, Bremen, Germany).  Repeated measurements with laboratory CO₂-in-air working standards had a precision of <0.1‰. The δ¹³C values of CO₂ in respiration samples were corrected to the international standards, Vienna Pee Dee Belemnite for ¹³C and the atmospheric nitrogen for ¹⁵N.  We calculated δ values (per mil, ‰) using δX = [(R_sample/R_standard) - 1] where X is the element of interest (¹³C or ¹⁵N), R_sample is for the ratio of the heavier to lighter isotope (¹³C/¹²C or ¹⁵N/¹⁴N) of the sample, and R_standard is for the ratio (¹³C/¹²C or ¹⁵N/¹⁴N) of the international standard. Reporting of isotope data and measurement results follow the guidelines provided by Coplen (2011).

Tissue and fractionation adjustments

We post-processed laboratory results of the source diet item in two ways.  First, we combined mammals into a single ‘mammal carrion’ category because: (1) we lacked sufficient species-specific sample sizes to estimate diet contribution of each mammal species, (2) all mammals shared similar ecological niches compared to other potential diet sources, and (3) ravens likely feed on mammal carcasses opportunistically.  We did not consider mammalian predators as potential diet items because Kristan et al. (2004) found evidence of predators in only 0.2% of raven pellets.  Second, we used data from the peer-reviewed literature to adjust the stable isotope values from our tissue samples to reflect fractionation within the body of each source species.  We did this because adult ravens were likely bringing energy-rich tissues (e.g., muscle) of mammals to the nest rather than our sampled tissue (e.g., fur).  We subtracted 1.6‰ from the δ¹³C value for all mammal fur samples to reflect mammal muscle (hereafter ‘mammal carrion’; Roth and Hobson 2000, Caut et al. 2009).  We subtracted 4.15‰ from the δ¹⁵N value and subtracted 1.85‰ from the δ¹³C value for all human hair samples to reflect human diet in the study area (hereafter ‘anthropogenic food’; Schoeller 1986, Minagawa 1992).  We subtracted 3.41‰ from the δ¹³C value for all sage-grouse egg membranes to reflect egg yolk and albumen but did not adjust the δ¹⁵N values for eggs (Hobson 1995).  We subtracted 2.14‰ from the δ¹⁵N value and 1.24‰ from the δ¹³C value for all sage-grouse feather samples to reflect sage-grouse muscle (hereafter ‘sage-grouse chick’).  We used previously published data from the literature to define the isotopic values of sagebrush steppe C₃ plants (δ¹⁵N = 1.54‰, δ¹³C = -26.60‰; Kelly 2000, Mowat and Heard 2006, Feranec 2007, Kohn 2010) and sagebrush steppe insects during spring (δ¹⁵N = 5.65‰, δ¹³C = -25.51‰; Blomberg et al. 2013).