Linking Precipitation and C3 - C4 Plant Production to Resource Dynamics in Higher Trophic Level Consumers: Insect Data (2005-2006)


In many ecosystems, seasonal shifts in temperature and precipitation induce pulses of primary productivity that vary in phenology, abundance and nutritional quality.  Variation in these resource pulses could strongly influence community composition and ecosystem function, because these pervasive bottom-up forces play a primary role in determining the biomass, life cycles and interactions of organisms across trophic levels.  The focus of this research is to understand how consumers across trophic levels alter resource use and assimilation over seasonal and inter-annual timescales in response to climatically driven changes in pulses of primary productivity. We measured the carbon isotope ratios (d13C) of plant, arthropod, and lizard tissues in the northern Chihuahuan Desert to quantify the relative importance of primary production from plants using C3 and C4 photosynthesis for consumers.  Summer monsoonal rains on the Sevilleta LTER in New Mexico support a pulse of C4 plant production that have tissue d13C values distinct from C3 plants.  During a year when precipitation patterns were relatively normal, d13C measurements showed that consumers used and assimilated significantly more C4 derived carbon over the course of a summer; tracking the seasonal increase in abundance of C4 plants.  In the following spring, after a failure in winter precipitation and the associated failure of spring C3 plant growth, consumers showed elevated assimilation of C4 derived carbon relative to a normal rainfall regime. These findings provide insight into how climate, pulsed resources and temporal trophic dynamics may interact to shape semi-arid grasslands such as the Chihuahuan Desert in the present and future.

Data set ID: 


Additional Project roles: 


Core Areas: 



Study site: 

This research was conducted on the Sevilleta LTER, located 100 km south of Albuquerque, New Mexico, which is an ecotonal landscape of Chihuahuan desert shrub and grasslands (Muldavin et al. 2008).  Data were collected from a 0.9 x 0.5km strip of land that encompassed a flat bajada and a shallow rocky canyon of mixed desert shrub and grassland dominated by the creosote bush (Larrea tridentata) and black grama grass (Bouteloua eriopoda). 

Tissue collection & sample preparation for stable isotope analysis:

From May to October of 2005 and 2006 we collected plant, lizard, and arthropod tissues for carbon stable isotope analysis. During mid-summer of 2005, we randomly collected leaf and stem samples from the 38 most abundant species of plants; these species produce over 90% of the annual biomass on our study site (see Appendix Table A).  Approximately 3.5 mg of plant material was then loaded into pre-cleaned tin capsules for isotope analysis.  

All animal research was conducted with the approval of the institutional animal care and use committee (UNM-IACUC #05MCC004).  Lizards were captured by hand using noose poles and by drift fence and pitfall trap arrays (Enge 2001) randomly scattered over a 0.5 km2 area.   Each lizard was toe clipped for permanent identification and snout-vent length (SVL), body mass (g) and sex were recorded.  For stable isotope analysis, we obtained a 50 μL blood sample from each lizard and only sampled individuals once in a two week period.  We acquired a total of 367 blood samples from 11 lizard species.  Blood samples were obtained by slipping a micro-capillary tube (Fisherbrand heparinized 50μL capillary tubes) ventral and posterior to the eyeball to puncture the retro-orbital sinus.   Before and after this procedure a local anesthesia (0.5% tetracaine hydrochloride ophthalmic solution, Akorn Inc.) was applied to the eye.  Blood samples were stored on ice and centrifuged within 24 hours to separate plasma and red blood cells.  For isotope analysis 15 μL of plasma were pipetted into a tin capsule, air dried, and then folded.  

Arthropods were captured bi-weekly from May through October of each year in pitfall traps (see above), as well as by hand and sweep netting. Individuals were frozen, lyophilized, ground into a fine powder and 0.5 mg samples were loaded into tin capsules for isotope analysis.

Stable isotope analysis:

Carbon isotope ratios of samples were measured on a continuous flow isotope ratio mass spectrometer (Thermo-Finnigan IRMS Delta Plus) with samples combusted in a Costech ECS 4010 Elemental Analyzer in the UNM Earth and Planetary Sciences Mass Spectrometry lab.  The precision of these analyses was ± 0.1‰ SD for δ13C.  A laboratory standard calibrated against international standards (valine δ13C -26.3‰ VPDB [Vienna Pee Dee Belemnite Standard]) was included on each run in order to make corrections to raw values. Stable isotope ratios are expressed using standard delta notation (δ) in parts per thousand (‰) as: δX = (Rsample /Rstandard – 1) x 1000, where Rsample and Rstandard are the molar ratios of 13C/12C of a sample and standard. 

Estimation of C3 and C4 carbon incorporation into arthropods and lizards:

We used d13C values of consumer tissues and a two-end-point mixing model to estimate the proportion of a consumer’s assimilated carbon that was derived from each plant photosynthetic type (Martinez del Rio and Wolf 2005):  

In this model p is the fraction of dietary C4 plant material incorporated into a sampled tissue. We chose to analyze the isotope composition of whole bodies for arthropods because this best reflects the diet of lizards.  For lizards we chose plasma because it has a rapid 13C turnover rate with an inter-specific retention time ranging from 25 to 44 days (Warne et al. 2009b). In the above model Δ is a discrimination factor, which is defined as the difference in isotope values between an animal’s tissues and food when feeding on an isotopically pure diet (DeNiro and Epstein 1978).  For our mixing model estimates we used discrimination (Δ13C) values resulting from a diet switch study for two species of lizards (Sceloporus undulatus, and Crotaphytus collaris) fed a diet of C4 raised crickets (Warne et al. 2009b).  We found the plasma of these lizards had a mean Δ13C = -0.2 ± 0.4‰ VPDB, while crickets fed a C4 based dog food had a Δ13C = -0.9 ± 0.4‰.  Reviews of stable isotope ecology have reported Δ13C values for arthropods ranging from -0.5 ± 0.3‰ (Spence and Rosenheim 2005) to 0.3 ± 0.1‰ (McCutchan et al. 2003).  Although variation in our assumed Δ13C values would affect proportional estimates of the C3 or C4 resources consumed, the observed trends would not change. 

Data analysis:  

To compare the seasonal isotope values of consumers between a spring C3 dominated and a summer C4 dominated ecosystem we present the mean δ13C (± SE) of each consumer species during the pre-monsoon (May, June and early-mid-July) and monsoonal periods for each year of this study.   We defined the monsoon period to begin with the first day of recorded monsoon rains in July (monsoon 2005 = July 25 to October 15; monsoon 2006 = July 6 to October 15).  The effects of seasonal and inter-annual primary production patterns on consumer resource assimilation (δ13C) were determined by two-way ANOVA using the PROC MIXED procedure (Littell et al. 2006) in SAS (SAS 1999).  To examine these effects in the lizard community as a whole, lizard species were treated as random effects in the PROC MIXED model.  In order to determine the significance of seasonal and year effects post-hoc analyses were conducted using Tukey-Kramer’s hsd test (Sokal and Rohlf 1995).  Prior to analysis the data were tested for homogeneity of variance and confirmed to meet the assumptions of ANOVA.

Data sources: 


Scaling of Recruitment with Seed Distribution and Colony Size in Pogonomyrmex spp. at the Sevilleta National Wildlife Refuge, New Mexico


Ant colonies possess a “societal metabolism,” acquiring, transforming, and allocating resources through a network of foragers (Moses, 2005). Ant foraging- trail networks channel foragers to known food resources and away from competing colonies (Jun et al., 2003). Computer models suggest the spread of information occurs faster in larger colonies of harvester ants, genus Pogonomyrmex (Adler and Gordon, 1992), providing a possible mechanism of differentiation. Does the ability to utilize and share information scale super-linearly with a colony’s size? Within colonies, do foragers recruit more to denser sources of food, using information transfer to increase forager efficiency and harvest seed caches before competing colonies find them? To address these questions, we studied three sympatric species of Pogonomyrmex in central New Mexico that differ in average colony size: P. rugosus, P. maricopa, and P. desertorum. We hypothesized a) that across colonies recruitment to dense food resources scales positively with colony size, and b) that within colonies recruitment scales positively with seed density. We observed baited colonies for 1 hr, tracking the capture of dyed seeds arranged in piles of different densities and of native seeds. We generated a model of idealized effects of recruitment on foraging patterns and compared the output to our observations. We did not find support for hypothesis a, that recruitment scales positively with colony size, but did find support for hypothesis b, that recruitment does scale positively with increasing seed density. These findings highlight a key intersection between the metabolism of energy and of information.

Core Areas: 

Data set ID: 




To explore whether large ant colonies are "smarter" and how colonies use recruitment.


Experimental Design: Our experimental design consisted of testing potential differences in forager recruitment by baiting actively foraging colonies with dyed seeds arranged in different distributions (i.e. piles of different numbers of seeds) around each colony. Differences in the rate foragers collect seeds of different colors indicate potential differences in the rate foragers are recruited, by chemical pheromone trails, to piles of different sizes. We conduct these experiments on a number of colonies of three species of Pogonomyrmex that differ in average colony size, and then compare the rates those species retrieve seeds of different distributions.

Field Methods: We located actively foraging Pogonomyrmex colonies between 8 and 9am and distributed dyed millet seeds in a wide circular swath around each colony. We used 256 millet seeds of each color for P. rugosus and P. maricopy and 32 sesame seeds of each color for P. desertorum, which is smaller and has smaller colonies. For each color, we divided those seeds into piles as follows: red = 1 pile; purple = 4 piles; green = 16 piles; blue = "piles" of one seed each (i.e. scattered at random within the circular swath). We then sat near the nest entrance and recorded the seeds brought in by foraging ants. Using a Java program (SeedCounter) we recorded the color and time of each seed retrieved by the focal colony during an observation period of 1 to 1.5 hours. We generated seed uptake curves from this data.

Laboratory Procedures: We also developed an ant foraging simulation. Virtual ants start at a nest entrance at the center of a bounded lattice. The lattice contains 256 randomly distributed blue seeds and a single pile of 256 red seeds. A foraging ant searches at random until it collects a seed, then delivers it to the nest. To model patch fidelity, a successful forager returns to location it found its previous seed and begins a new random search.

Keystone species have large impacts on community and ecosystem properties, and create important ecological interactions with other species.  Prairie dogs (Cynomys spp.) and banner-tailed kangaroo rats (Dipodomys spectabilis) are considered keystone species of grassland ecosystems, and create a mosaic of unique habitats on the landscape.

Subscribe to RSS - invertebrates