Undecane detection
Alexandra Koch
15 3 2021
Background and Methods
We aimed to determine a solution of alarm pheromone needed to elicit a sufficient but realistic aggressive reaction in Lasius niger ants.
To do so, we conducted a serial dilution of undecane in dichloromethane (DCM) in the order of magnitudes and then conducted a blinded bioassay.
Methods
For the behavioural assay we used commercially available 0.2 l plastic cups. In each cup a subgroup of five ants of one colony was tested per treatment (i.e. one cup = one observation). The rim of each plastic cup was coated with fluon to prevent ants from escaping. We used small pieces of paper of ~ 5 mm Ø as carrier medium for the respective treatment. Paper pieces were made by punching paper with a hole punch. We tested six different treatments: four dilutions, pure DCM as control for the solvent and plain paper piece as control for the handling of the ants (see Analysis). Treatments were blinded, so that the experimenter did not know which dilution was tested, except for the plain paper control treatment. After placing the ants in each cup, ants were allowed to rest and settle for ≥ 30 min before testing. Testing was conducted by pipetting 0.1 µl of the first treatment liquid (except for plain paper control) onto the paper piece. The liquid on the paper was allowed to evaporate for ~ 30 s. The paper piece was then placed with tweezers in the centre of the respective plastic cup bottom without any direct contact to the cup. Two cups (i.e. two ant groups) were tested simultaneously. The behaviour of the each group was observed for 2 min. Behavioural reactions to the treatment were noted as described in the table below. Treatments were tested sequentially with 15 min breaks between treatments to minimize possible airborne interference of prior treatment, e.g. each colony was first tested with treatment A, followed by a 15 min break and then the next cups were tested with treatment B.
| Behaviour | Description |
|---|---|
| Antennation | Do any of the ants lift/stretch their antenna towards the piece of paper (yes or no) |
| Movement | Do any of the ants move towards the piece of paper and directly touch it with their antenna (yes or no) |
| Mandible opening | Do any of the ants open their mandibles in a defensive manner (yes or no) |
| Amount of ants opening mandibles | How many ants of the group open their mandibles in a defensive manner? |
| Attacking of paper | Do any of the ants attack the piece of paper, i.e. quickly runs towards the paper with open mandibles or bites the paper (yes or no) |
| Grooming | After having touched the paper with the antenna, do any of the ants groom (yes or no) |
If an ant showed a given reaction in an observation (each observation is 1 cup with 5 ants), it was noted/ scored as 1. Except for the amount of ants opening mandibles.
Analysis
Read and clean data
Read in raw data (tested blind from Monday to Wednesday 15th - 17th March).
Rename control treatment using plain paper to single letter (“P” for paper) and check number of observations per treatment.
##
## A B C D E P
## 26 26 26 26 26 26| Treatment | Concentration of Undecane |
|---|---|
| A | 0.0001 % |
| B | 0 % (i.e. pure DCM as solvent control) |
| C | 0.01 % |
| D | 0.001 % |
| E | 0.00001 |
| P | control with plain paper piece (no solution/solvent on paper) |
Concentration of undecane refers to Undecane in 1 µl Undecane-DCM-solution.
Reactions of ants to each treatment was analysed using a reaction score. The reaction score was created by summing the reactions (antennation, movement, mandible opening, amount of ants opening mandibles, attacking paper and grooming), thus, the amount of ants opening mandibles had a stronger weight on the score. We assume the higher the reaction score (the more reaction ants showed) in one treatment the more likely ants reacted (aggressively) towards undecane.
## [1] "0" "0.0001" "0.01" "0.001" "0.00001" "control"## [1] "0" "0.00001" "0.0001" "0.001" "0.01" "control"Plots
Reaction score
fit <- glmer(reaction_score ~ treatment +
(1|cup_ID),
data = raw.dat,
family = "poisson")
res.fit <- simulateResiduals(fit)
plot(res.fit)
plot(allEffects(fit))
summary(fit)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: reaction_score ~ treatment + (1 | cup_ID)
## Data: raw.dat
##
## AIC BIC logLik deviance df.resid
## 622.0 643.4 -304.0 608.0 149
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.49832 -0.65803 0.03499 0.37941 2.05916
##
## Random effects:
## Groups Name Variance Std.Dev.
## cup_ID (Intercept) 0.2382 0.488
## Number of obs: 156, groups: cup_ID, 144
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.69722 0.07925 8.798 < 2e-16 ***
## treatment.L 0.46610 0.16423 2.838 0.00454 **
## treatment.Q 0.34404 0.16809 2.047 0.04068 *
## treatment.C 0.11264 0.16878 0.667 0.50455
## treatment^4 -0.02171 0.17118 -0.127 0.89907
## treatment^5 0.29568 0.17658 1.675 0.09403 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) trtm.L trtm.Q trtm.C trtm^4
## treatment.L -0.114
## treatment.Q -0.079 -0.099
## treatment.C -0.026 -0.047 -0.109
## treatment^4 0.015 -0.048 -0.052 0.001
## treatment^5 -0.109 0.013 0.066 -0.038 -0.072Amount of mandibles opened
fit2 <- glmer(amount_mandibles ~ treatment +
(1|cup_ID),
data = raw.dat,
family = "poisson",
control=glmerControl(optimizer="bobyqa",
optCtrl=list(maxfun=2e5)))
res.fit2 <- simulateResiduals(fit2)
plot(res.fit2)
plot(allEffects(fit2))
summary(fit2)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: amount_mandibles ~ treatment + (1 | cup_ID)
## Data: raw.dat
## Control:
## glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+05))
##
## AIC BIC logLik deviance df.resid
## 253.3 274.6 -119.6 239.3 149
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.0305 -0.4344 -0.4038 0.2750 3.4744
##
## Random effects:
## Groups Name Variance Std.Dev.
## cup_ID (Intercept) 0.5316 0.7291
## Number of obs: 156, groups: cup_ID, 144
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.404478 0.257484 -5.455 4.91e-08 ***
## treatment.L 0.977251 0.361708 2.702 0.0069 **
## treatment.Q 0.887141 0.409103 2.169 0.0301 *
## treatment.C -0.003646 0.389720 -0.009 0.9925
## treatment^4 -0.413943 0.407661 -1.015 0.3099
## treatment^5 0.163810 0.466963 0.351 0.7257
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) trtm.L trtm.Q trtm.C trtm^4
## treatment.L -0.232
## treatment.Q -0.194 -0.170
## treatment.C 0.024 -0.100 -0.222
## treatment^4 0.139 -0.017 -0.220 -0.104
## treatment^5 -0.076 0.084 0.097 -0.211 -0.177Posthoc analysis
em <- emmeans(fit, specs = pairwise ~ treatment)
print(em)## $emmeans
## treatment emmean SE df asymp.LCL asymp.UCL
## control 0.542 0.182 Inf 0.1858 0.897
## 0 0.657 0.174 Inf 0.3163 0.997
## 0.00001 0.330 0.195 Inf -0.0521 0.713
## 0.0001 0.747 0.169 Inf 0.4168 1.078
## 0.001 0.687 0.174 Inf 0.3464 1.028
## 0.01 1.220 0.148 Inf 0.9295 1.510
##
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df z.ratio p.value
## control - 0 -0.1151 0.245 Inf -0.469 0.9972
## control - 0.00001 0.2112 0.258 Inf 0.818 0.9644
## control - 0.0001 -0.2056 0.242 Inf -0.849 0.9582
## control - 0.001 -0.1456 0.244 Inf -0.597 0.9912
## control - 0.01 -0.6783 0.229 Inf -2.963 0.0361
## 0 - 0.00001 0.3263 0.255 Inf 1.281 0.7956
## 0 - 0.0001 -0.0905 0.238 Inf -0.381 0.9990
## 0 - 0.001 -0.0305 0.240 Inf -0.127 1.0000
## 0 - 0.01 -0.5632 0.224 Inf -2.514 0.1199
## 0.00001 - 0.0001 -0.4168 0.252 Inf -1.655 0.5620
## 0.00001 - 0.001 -0.3567 0.253 Inf -1.410 0.7209
## 0.00001 - 0.01 -0.8895 0.239 Inf -3.718 0.0028
## 0.0001 - 0.001 0.0600 0.237 Inf 0.254 0.9999
## 0.0001 - 0.01 -0.4727 0.220 Inf -2.144 0.2646
## 0.001 - 0.01 -0.5327 0.223 Inf -2.390 0.1597
##
## Results are given on the log (not the response) scale.
## P value adjustment: tukey method for comparing a family of 6 estimatesplot(em)
em2<- emmeans(fit2, specs = pairwise ~ treatment)
print(em2)## $emmeans
## treatment emmean SE df asymp.LCL asymp.UCL
## control -1.592 0.467 Inf -2.51 -0.677
## 0 -1.567 0.456 Inf -2.46 -0.673
## 0.00001 -2.169 0.574 Inf -3.29 -1.044
## 0.0001 -1.727 0.485 Inf -2.68 -0.777
## 0.001 -0.966 0.375 Inf -1.70 -0.232
## 0.01 -0.406 0.318 Inf -1.03 0.217
##
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df z.ratio p.value
## control - 0 -0.0244 0.586 Inf -0.042 1.0000
## control - 0.00001 0.5775 0.675 Inf 0.856 0.9567
## control - 0.0001 0.1353 0.607 Inf 0.223 0.9999
## control - 0.001 -0.6258 0.529 Inf -1.184 0.8446
## control - 0.01 -1.1860 0.497 Inf -2.387 0.1606
## 0 - 0.00001 0.6019 0.674 Inf 0.893 0.9482
## 0 - 0.0001 0.1597 0.605 Inf 0.264 0.9998
## 0 - 0.001 -0.6014 0.526 Inf -1.143 0.8634
## 0 - 0.01 -1.1616 0.494 Inf -2.353 0.1733
## 0.00001 - 0.0001 -0.4422 0.691 Inf -0.640 0.9880
## 0.00001 - 0.001 -1.2033 0.624 Inf -1.929 0.3845
## 0.00001 - 0.01 -1.7634 0.597 Inf -2.952 0.0373
## 0.0001 - 0.001 -0.7612 0.549 Inf -1.387 0.7351
## 0.0001 - 0.01 -1.3213 0.518 Inf -2.551 0.1097
## 0.001 - 0.01 -0.5601 0.422 Inf -1.326 0.7708
##
## Results are given on the log (not the response) scale.
## P value adjustment: tukey method for comparing a family of 6 estimatesplot(em2)
Summary
| Treatment | Concentration of Undecane |
|---|---|
| A | 0.0001 % |
| B | 0 % (i.e. pure DCM as solvent control) |
| C | 0.01 % |
| D | 0.001 % |
| E | 0.00001 |
| P | plain paper piece (no solution/solvent on paper) |
Concentration of undecane refers to undecane in 1 µl undecane-DCM-solution.
Results show that treatment significantly affects the aggressive reaction of the ants. Reaction score, that is the sum of all behavioural events noted during each observation period was highest in treatment C, i.e. with a 0.01 % concentration of undecane in DCM. Treatment C showed a significantly higher reaction score than treatment E or the plain paper control. Reaction scores to treatments with lower concentrations of undecane than treatment C (i.e. A, D and E) did not significantly differ from the control condicitons (solvent control and plain paper control).
Due to these results one can assume that a solution with a concentration of 0.01 % Undecane in DCM can elicit a sufficient aggressive behaviour in Lasius niger.