Incremental increases in a driver variable, such as nutrients or detritus, can trigger abrupt shifts in aquatic ecosystems that may exhibit hysteretic dynamics and a slow return to the initial state. A model system for understanding these dynamics is the microbial assemblage that inhabits the cup-shaped leaves of the pitcher plant Sarracenia purpurea. With enrichment of organic matter, this system flips within three days from an oxygen-rich state to an oxygen-poor state. In a replicated greenhouse experiment, we enriched pitcher-plant leaves at different rates with bovine serum albumin (BSA), a molecular substitute for detritus. Changes in dissolved oxygen (DO) and undigested BSA concentration were monitored during enrichment and recovery phases. With increasing enrichment rates, the dynamics ranged from clockwise hysteresis (low), to environmental tracking (medium), to novel counter-clockwise hysteresis (high). These experiments demonstrate that detrital enrichment rate can modulate a diversity of hysteretic responses within a single aquatic ecosystem, and suggest different management strategies may be needed to mitigate the effects of high versus low rates of detrital enrichment.

Replicate S. purpurea pitchers were enriched with 5.0 (high concentration) or 0.5 (low concentration) mg of bovine serum albumi (BSA) per mL of pitcher fluid in 2015 and 2.0 mg/mL (intermediate concentration) in 2016. For each experiment, a set of control pitchers received no organic matter. These controls were used to determine when oxygen levels in experimental pitchers had recovered. Experimental pitchers were loaded with organic matter (between 9:00 am and 9:45 am) following pitcher-fluid sampling and DO measurement for the first four days of the experiment. 

We measured DO twice a day each day from day 20 to day 28, and once on Days 30, 31, 33, and 35. DO was measured using a D-166MT-1S microelectrode (Lazar Research Laboratories). The microelectrode was calibrated prior to each sampling event according to the manufacturer’s instructions. A single microelectrode was used to take readings from each pitcher and was rinsed twice with reverse osmosis water between readings. The order of readings from different replicates was randomized so that changes in temperature and sunlight over the sampling period were not confounding factors. During readings, the microelectrode was placed 2.5 cm below the surface of the pitcher fluid and swirled so that the more oxygen-rich pitcher fluid at the top of the pitcher was mixed and readings reflected average DO. Due to the sensitivity of the microelectrode, readings were taken as soon as the reader settled on a value for more than 10 s. 

Pitcher fluid was sampled following each DO measurement for spectrophotometric analysis using a Bradford assay. Using sterile pipette tips, a 300 µL aliquot was taken from 2.5 cm below the surface of each pitcher and placed in a sterile 1-mL microfuge tube. Sample tubes were immediately transported to the lab where they were centrifuged at 13,000 xg for two minutes. The supernatant containing soluble BSA was removed, placed in a sterile 1-mL microfuge tube, and stored at -80°C until analyzed.

The data were analyzed using the following R scripts: HYS_vFinal_removedreplicates, hysteresis_functions.R, reviewer3figure.R

This dataset was collected in two separate years (2015 and 2016). High and low concentration treatment data were collected in 2015 and interediate treatment data were collected in 2016.