In this experiment, efficacy of minibucket traps baited with cuelure and UVLEDs was compared to that of pan traps baited likewise. Despite the fact that the pan traps have a known escape rate of about 25%, these traps outperformed minibuckets by more than an order of magnitude. Median trap catch was significantly greater for pan traps compared to minibuckets traps, 0.14 and 0.00 beetles per trap day respectively, p = 3.2e-10, Mann-Whitney Rank Sum Test. Barrel traps caught an average of 0.263 beetles per trap-day, while minibucket traps caught only 0.016 beetles per trap-day, which is approximately equivalent to the trap catch rate for standard CRB pheromone traps. Note that the barrels supporting the pan and minibucket traps were empty: they were not baited with breeding site material.
Findings from this experiment suggest that the minibucket trap design should be abandoned and the pan trap design should be further investigated with emphasis on reducing the escape rate to near zero. Possibly, this can be acheived using baffles to prevent beetles from flying up to the chicken wire.
In previous experiments, CRB2014-01-12A and CRB2014-01-15, we determined that about 25% of beetles escape through the chicken wire after entering pan traps. As a possible replacment for the pan traps, we built and tested minibucket traps, similar in design to traps used by SPC. The escape rate for these traps is near zero, CRB2014-01-17, and the trap caught beetles in a large field cage test, CRB2014-01-16. This experiment is the first field trial for the minibucket trap on Guam.
Thirty-one empty oil barrels were placed on the University of Guam Agricultural Experiment Station in Yigo with a inter-barrel spacing of about 100 m. Alternate barrels supported pan traps and minibucket traps. The experiment ran for 3 weekly trapping periods. At the start of each trapping period, pan traps were exchanged with minibucket traps.
%pylab inline
%load_ext sql
%sql mysql+pymysql://readonlyguest:readonlypassword@mysql.guaminsects.net/oryctes
%sql SELECT * FROM YigoBarrelObs LIMIT 1
Populating the interactive namespace from numpy and matplotlib 1 rows affected.
barrelID | trapType | lure | uvled | substrate | startDate | endDate | males | females | notes |
---|---|---|---|---|---|---|---|---|---|
1 | P | y | y | n | 2014-05-12 | 2014-05-19 | 1 | 2 | lure not hung |
# Get data
P = %%sql SELECT (males+females)/(endDate-startDate) AS beetlesPerTrapDay FROM YigoBarrelObs WHERE trapType = 'P' AND endDate <= '2014-06-02'
B = %sql SELECT (males+females)/(endDate-startDate) AS beetlesPerTrapDay FROM YigoBarrelObs WHERE trapType = 'B' AND endDate <= '2014-06-02'
import pandas as pd
Pdf = pd.DataFrame(P, columns=P.keys)
Bdf = pd.DataFrame(B, columns=B.keys)
# Plot
figure(figsize=(10,5))
plt.boxplot([Pdf, Bdf])
plt.ylim(-.1,2)
plt.xlim(0, 2.5)
plt.xticks([1,2],['pan','minibucket'])
plt.ylabel('beetles per trap-day')
# Overplot raw data with jitter on x values
x = numpy.random.uniform(0.5, 0.8, len(Pdf.index))
plt.plot( x, Pdf['beetlesPerTrapDay'], 'ro')
x = numpy.random.uniform(1.5, 1.8, len(Bdf.index))
sponge = plt.plot( x, Bdf['beetlesPerTrapDay'], 'ro')
# Stats
from scipy import stats
z_stat, p_val = stats.ranksums(Pdf, Bdf)
print 'Mean beetles per trap-day for minibuckets = ', Bdf['beetlesPerTrapDay'].mean()
print 'Median beetles per trap-day for pans = ', Bdf['beetlesPerTrapDay'].median()
print 'Mean beetles per trap-day for minibuckets = ', Pdf['beetlesPerTrapDay'].mean()
print 'Median beetles per trap-day for pans = ', Pdf['beetlesPerTrapDay'].median()
print "MWW RankSum P for beetles per trap-day for pans and minibuckets =", p_val
47 rows affected. 46 rows affected. Mean beetles per trap-day for minibuckets = 0.0155279503106 Median beetles per trap-day for pans = 0.0 Mean beetles per trap-day for minibuckets = 0.263302449921 Median beetles per trap-day for pans = 0.142857142857 MWW RankSum P for beetles per trap-day for pans and minibuckets = 3.1844810952e-10