Electronic Supplementary Material (ESM): Online Resource 1
Online Resource 1. Summary of experimental manipulations of prey for field mesocosm experiment, and arena area for laboratory experiment.
Predator interference alters foraging behavior of a generalist predatory arthropod
List of Authors. –
Jason M. Schmidt1, *, Thomas O. Crist1, Kerri Wrinn1 and Ann L. Rypstra2
1 Department of Zoology, Miami University, Oxford, OH, 45056, USA
2Department of Zoology, Miami University, Hamilton, OH, 45011, USA
*Corresponding author present address and contact information:
Jason M Schmidt PhD
Assistant Professor |Biocontrol lab
UGA Entomology |Coastal Plain Experimental Station
2360 Rainwater Rd,Tifton, GA 31793
Ph:(229) 386-7251|Email:
Total number of prey added organized by prey order.Seven 1.5m2 samples were suction sampled from the grassy border surrounding the soybean fields located at the Ecology Research Center, Butler County.This addition of prey to mesocosms was approximately an increase in density of prey by 2316m-2.Standard deviation was calculated on the mean number of insects of the seven samples.
Prey order / Number of prey added / Std. dev.Coleoptera / 11 / 1.17
Collembola / 1235 / 167.62
Diptera / 135 / 7.87
Hemiptera / 19 / 2.04
Homoptera / 335 / 18.56
Orthopera / 2 / 0.52
Description of procedure.
We used a caged laboratory approach to estimate the magnitude of interference between these predators. The table below represents container sizes used for different densities to approximately hold the search surface area each spider, Pardosamilvina, was exposed to constant.Pardosaare numerically abundant and can reach 50 spiders m-2 when trophic resources and habitat structure are provided (Marshallet al. 2002).Furthermore, spiders in the genus Pardosa occur at much higher densities (600 spiders m-2) in other systems (Finke andDenno 2002). In addition, because we provided preferred habitat structure (Schmidt andRypstra 2010), this makes the experimental conditions more realistic in light of being conducted in small laboratory containers.
Spider abundance treatment / Diameter of container (cm) / Area to spider ratio (cm2/spider)1 / 9.5 / 70.8
2 / 14.0 / 76.9
4 / 20.5 / 82.4
8 / 25.5 / 63.8
References
Finke DL, Denno RF (2002) Intraguild predation diminished in complex-structured vegetation: Implications for prey suppression. Ecology83: 643–652.
MarshallS, Pavuk DandRypstra AL (2002) A comparative study of phenology and daily activity patterns in the wolf spiders Pardosamilvina and Hognahelluo in soybean agroecosystems in southwestern Ohio (Araneae, Lycosidae). Journal of Arachnology30: 503–510.
Schmidt JMandRypstra AL (2010) Opportunistic predator prefers habitat complexity that exposes prey while reducing cannibalism and intraguild encounters. Oecologia164: 899–910.
Electronic Supplementary Material (ESM): Online Resource 2
Online Resource 2. Description of R code programming used to estimate functional response parameters.
Predator interference alters foraging behavior of a generalist predatory arthropod
List of Authors. –
Jason M. Schmidt1, *, Thomas O. Crist1, Kerri Wrinn1 and Ann L. Rypstra2
1 Department of Zoology, Miami University, Oxford, OH, 45056, USA
2Department of Zoology, Miami University, Hamilton, OH, 45011, USA
*Corresponding author present address and contact information:
Jason M Schmidt PhD
Assistant Professor |Biocontrol lab
UGA Entomology |Coastal Plain Experimental Station
2360 Rainwater Rd,Tifton, GA 31793
Ph:(229) 386-7251|Email:
# Use numerical integration of instantaneous functional responses (fun.resp)
# and maximum likelihood estimation of sums of squares (sum.sq)
# p=vector of parameters: a=attack rate, b/a=handling time, m=interference
# constant (=c in BD and CM models)
# y=state variable (initial prey density)
# v=initial prey density of experimental treatment
# N=prey densities at time t
# np=number of predators; set to np=1 if there is a single predator
# T=total number of hours of functional response experiment
# n=number of observations
# lss=FALSE is sums of squares, and lss=TRUE is the log sums of squares
# fr=functional response type: "LV"=Lotka-Volterra,"HO"=Holling Type II,
# "HV=Hassell-Varley","AA"=Arditi-Akcakaya,"BD"=Beddington-DeAngelis",
# "CM"=Crowley-Martin
# the two functions sum.sq and fun.resp requirebblme and deSolve packages
library(bbmle)
library(deSolve)
sum.sq<-function(p,v,e,np) {
a<-p[1]
if (a<0) return(NA)
fun.resp<-function(y,t,p) {
with(as.list(p), {
y<-v
N<-y
if (fr=="LV") {
dN<- -a*N*np # Lotka-Volterra
}
if (fr=="HO") {
dN<- -a*N/(1+b*N)*np # Holling
b<-p[2]
}
if (fr=="HV") {
dN<- -a*N/(np^m)*np # Hassell-Varley
m<-p[2]
}
if (fr=="AA") {
dN<- -a*N/(b*N+np^m)*np # Arditi-Akcakaya
b<-p[2]
m<-p[3]
}
if (fr=="BD") {
dN<- -a*N/(1+b*N+m*(np-1))*np # Beddington-DeAngelis
b<-p[2]
m<-p[3]
}
if (fr=="CM") {
dN<- -a*N/((1+b*N)*(1+m*(np-1)))*np # Crowley-Martin
b<-p[2]
m<-p[3]
}
return(list(dN))
})
}
y0<-c(N=v)
T<-24
t<-0:T
out<-ode(y=y0,times=t,func=fun.resp,parms=p)
pred<-(out[1,2:(length(v)+1)]-out[(T+1),2:(length(v)+1)])/np
obs<-e/np
if (obs<=0 || pred<=0) return(NA)
if (lss)
ss<- sum((log(obs)-log(pred))^2)
else
ss<- sum((obs-pred)^2)
return(ss)
}
# Set options for analysis - Example using Arditi and Akcakaya
fr<-"AA"
lss<-FALSE
# Define parameters (a) for all models; (a,b) for HO, AA, BD, and CM;
# (a,m)for HV; and (a,b,m) for AA, BD, and CM
parnames(sum.sq)=c("a","b","m")
# Maximum likelihood estimation
mod1<-mle2(sum.sq,start=c(a=0.1,b=0.01,m=0.01),data=list(v=v,e=e,np=np))
print(fr)
summary(mod1)
AIC(logLik(mod1))
# Profile likelihoods
pr1<-profile(mod1)
plot(pr1)
# Example using Holling Type II
fr<-"HO"
parnames(sum.sq)=c("a","b")
mod2<-mle2(sum.sq,start=c(a=0.1,b=0.01),data=list(v=v,e=e,np=np))
print(fr)
summary(mod2)
AIC(logLik(mod2))
References
Beddington JR (1975) Mutual interference between parasites or predators and its effect on searching efficiency. Journal of Animal Ecology 45: 331–340.
Bolker B, R Development Core Team (2011) bbmle:Tools forGeneral Maximum Likelihood Estimation. R package version1.0.0.
Crowley PH, Martin EK (1989) Functional responses and interference within and between year classes of a dragonfly population. Journal of the North American Benthological Society8: 211–221.
DeAngelis DL, Goldstein RA, O’Neill RV (1975) A model for trophic interaction. Ecology56: 881–892.
Skalski GT, Gilliam JF (2001) Functional responses with predator interference: viable alternatives to the Holling Type II model. Ecology 82: 3083–3092.
Soetaert K, Petzoldt T, Woodrow S (2010) Solving Differential Equations in R: Package deSolve. Journal of Statistical Software33: 1–25.
Electronic Supplementary Material (ESM): Online Resource 3
Online Resource 3. Statistical results and additional data from analysis of field mesocosm study.
Predator interference alters foraging behavior of a generalist predatory arthropod
List of Authors. –
Jason M. Schmidt1, *, Thomas O. Crist1, Kerri Wrinn1 and Ann L. Rypstra2
1 Department of Zoology, Miami University, Oxford, OH, 45056, USA
2Department of Zoology, Miami University, Hamilton, OH, 45011, USA
*Corresponding author present address and contact information:
Jason M Schmidt PhD
Assistant Professor |Biocontrol lab
UGA Entomology |Coastal Plain Experimental Station
2360 Rainwater Rd,Tifton, GA 31793
Ph:(229) 386-7251|Email:
Description of procedure.
Body condition measurements for wolf spiders in the field experiment.Body condition was measured from seven wolf spiders, Pardosamilvina, from each treatment. Prey treatments indicate whether prey were removed or added prior to introduction of the spider abundance treatments. Spider abundance represent low = 7 spiders or high = 28 spiders.Spiders from the high-density treatment were randomly selected from the 28 spiders initially entered into the enclosures.
Prey treatment / Predator abundance treatment / Number of spiders measured / Carapace width mean (±1std) mm / Abdomen width mean (±1std) mmRemoved / Low / 56 / 2.13 (0.22) / 2.10 (0.39)
Removed / High / 56 / 2.14 (0.17) / 2.07 (0.41)
Added / Low / 56 / 2.09 (0.23) / 1.97 (0.38)
Added / High / 56 / 2.11 (0.19) / 2.03 (0.46)
Results from fitting a generalized linear model to the binomial response variable, proportion of Pardosamilvina leaving field based enclosures. Coefficients represent the intercept and predictor variables in the model. The prey availability treatments were either prey removed or added and predator abundance treatments were low and high abundance (7 and 28 spiders respectively).
Coefficients / Estimate / Std. Error / t value / PIntercept / -0.99 / 0.16 / -6.34 / <0.0001 / ***
Prey availability / 0.90 / 0.21 / 4.15 / 0.0001 / ***
Predator abundance / -0.93 / 0.43 / -2.15 / 0.031 / *
Interaction / 1.02 / 0.55 / 1.88 / 0.061
Analysis of fixed effects from a linear mixed modelanalysis of consumption of prey, change in abdomen width (mm), by Pardosamilvina in response to prey availability and predator abundancetreatments in field enclosures (n = 8 per treatment combination).
df / F-value / PIntercept / 1, 149 / 9.51 / 0.002
Prey treatment / 1, 23 / 13.22 / 0.001
Predator abundance / 1, 23 / 0.01 / 0.926
Interaction / 1, 23 / 1.45 / 0.241
Mean proportion ±1SE of spiders, Pardosamilvina, cannibalized (e.g. ratio of the number missing to recovered from within the enclosures or attached cups) at the end of the 24-hr experimental period as an estimate of the frequency of cannibalism. Enclosures were the experimental unit where both prey and spiders were introduced that contained holes drilled to allow for spiders to exit into attached pitfall cups to measure emigration.