44 Journal of Vector Ecology

June 2017

Effect of CO2 concentration on mosquito collection rate using odor-baited suction traps
Lee McPhatter* and Alec C. Gerry

Department of Entomology, University of California, Riverside, CA, U.S.A., [email protected]

Received 4 August 2016; Accepted 11 November 2016

ABSTRACT: Carbon dioxide (CO2) has been used for decades to enhance capture of host-seeking mosquitoes when released in
association with traps commonly used by mosquito and vector control agencies. However, there is little published work evaluating the
effect of altering CO2 release rates relative to the number of mosquitoes captured in these traps. This study investigated how varying
CO2 concentration altered the mosquito collection rate at a freshwater wetlands in southern California. Host-seeking mosquitoes were
captured in CDC-style traps baited with one of six CO2 release rates ranging from 0-1,495 ml/min from gas cylinders. Species captured
were Aedes vexans, Anopheles franciscanus, An. hermsi, Culex erythrothorax, and Cx. tarsalis. A biting midge, Culicoides sonorensis, was
also captured. For all species, increasing CO2 release rates resulted in increasing numbers of individual females captured, with the relative
magnitude of this increase associated to some extent with known feeding preferences of these species. We found that variation in CO2
release rate can significantly alter mosquito capture rates, potentially leading to imprecise estimates of vector activity if the relationship
of CO2 release rate to mosquito capture rate is not considered. Journal of Vector Ecology 42 (1): 44-50. 2017.

Keyword Index: Mosquito trap, Culex, Anopheles, Culicoides, host-seeking, surveillance.

INTRODUCTION type of holding container and environmental conditions in the

field.
Hematophagous insects rely on biochemical cues such Host CO2 output varies considerably across a range of
as odors to locate and identify hosts on which to feed. Carbon potential animal hosts. For example, Roberts (1972) found the
dioxide is one of the most important olfactory stimuli involved CO2 emissions of beef heifers ranged from 1,200 to 1,800 ml/
in orientation toward hosts by mosquitoes and other blood min, while Gillies and Wilkes (1974) determined CO2 output for
feeding insects (Gillies 1980). Respiration involving CO2 output chickens to be 50 ml/min. The CO2 output of a standard CDC-style
is common to all vertebrates and therefore serves as a signal for suction trap baited with dry ice held in an insulated container has
their presence. been recorded to range from 1,000-1,500 ml/min (Mullens 1995)
Surveillance and monitoring for changes in the abundance or 641-892 ml/min (Eshun et al. 2016), or roughly equivalent to a
and activity of host-seeking mosquitoes is critical to assess the large mammal.
public health risk for mosquito-transmitted pathogens like West Several investigators have documented changes in the capture
Nile virus (WNV). Surveillance should employ efficient collection rate of biting flies when releasing variable CO2 concentrations
methods that allow for definitive species identification while also (release rate or flow rate) to attract these flies to a suction trap.
providing an accurate representation of species host-seeking Carestia and Savage (1967) used traps with CO2 flow rates of 250,
activity. Numerous studies have shown that the release of CO2 will 500, 1,000, and 2,000 ml/min and reported that mosquito capture
enhance the capture of host-seeking mosquitoes when released generally increased with increasing CO2 release rate for a number
with traps commonly used by mosquito control agencies (e.g., of species (Anopheles quadrimaculatus Say, An. walkeri Theobald,
CDC light trap (Sudia and Chamberlain 1962) or Encephalitis An. punctipennis Say, Aedes vexans (Meigen), Culex pipiens L.,
Virus Surveillance trap (Rohe and Fall 1979)). However, there is and Cx. salinarius Coquillett), but traps in their study were
little published work analyzing the effect of altering CO2 release not run concurrently and thus could not be directly compared
rates relative to the number of mosquitoes captured in the field. for differences among the CO2 release rates. Additionally, the
Previous non-laboratory studies by Reeves (1953) and Reisen et relationship of CO2 release rate and insect capture rate may differ
al. (2000) focused primarily on Culex tarsalis Coquillett attraction among species, as later studies showed that increasing the CO2
to traps using measured CO2 release rates simulating specific hosts concentration released was not always associated with an increased
(chicken, human, cow) and in relation to varying trap design capture of host-seeking insects. For example, Pfuntner et al. (1988)
with CO2 presentation (e.g., pulsating vs consistent CO2 release), found that while capture of Cx. tarsalis increased significantly
respectively. when CO2 release was increased from 250 to 1,000 ml/min, the
There are three common sources for CO2 used in mosquito capture of Cx. quinquefasciatus Say and Cx. sitgmatosoma Dyar
traps: dry ice, compressed gas, and hydrocarbon fuel combustion. was not altered by the increased CO2 release. In contrast, Mullens
Dry ice is most commonly used for vector surveillance by and Gerry (1998) observed that capture of Cx. quinquefasciatus
mosquito control and public health agencies due to lower cost and actually decreased significantly when the CO2 release rate was
portability. However, the concentration of CO2 released by dry ice increased from 300 to 1,000 ml/min, while simultaneous capture
is highly variable, as the quantity and surface area (solid block of the highly mammalophilic biting midge Culicoides sonorensis
or pellet) of dry ice used with traps will often differ considerably Wirth & Jones increased with the higher CO2 release rate. For
among trap designs. Additionally, sublimation rates vary with the black flies identified as Cnephia (Stegopterna) mutata (Malloch)
 19487134, 2017, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jvec.12238 by CAPES, Wiley Online Library on [03/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Vol. 42, no. 1 Journal of Vector Ecology 45

and Prosimulium hirtipes (Fries), it was determined that CO2 with trap nights separated by 2-3 days within each week. Trapping
release rates exceeding 500 ml/min did not further increase their began 30 min after sunset and ended two h later to cover the
collection (Frommer et al. 1976). Both C. (Stegopterna) mutata peak mosquito activity period at this site based upon preliminary
and P. hirtipes are known to bite humans (Mokrey 1978, Mason studies. The 2-h trapping period also assured that CO2 flow rates
and Schmanchuck 1990) though C. mutate is also recorded to remained consistent over the entirety of the trapping period. Low
be autogenous and thus does not require an initial blood meal CO2 flow rates (15 and 47 ml/min) were tested on the first night
(Mokrey 1978). each week (low CO2 nights), while high CO2 flow rates (473 ml/
The objectives for this study were (1) to investigate how min and 1,495 ml/min) were tested on the second night each
varying CO2 concentration alters capture of mosquitoes at a week (high CO2 nights). High and low flow rates were separated
southern California wetlands, and (2) to determine if changes in by night to avoid masking or interference of traps with low CO2
the mosquito capture rate were associated with increasing CO2 flow rates by those with high CO2 flow rates. An intermediate or
concentration similarly for available host-seeking species at the mid-range flow rate (149 ml/min) and a negative control (0 ml/
wetlands site. min) were included on all nights in order to make comparisons
across collection nights. During each trap night, two traps were
MATERIALS AND METHODS set with each of the four flow rates to be tested on that night,
for a total of eight traps utilized per night. One trap with each
Host-seeking mosquitoes, and other hematophagous Diptera, flow rate was randomly assigned to the first four trap positions
were collected from 10 July 2013 through 28 August, 2013 at the of the circular trapline (Figure 1), and the second trap with the
San Jacinto Wildlife Area (SJWA; 33°52´14.19˝N, 117°7´6.86˝W). same flow rate was then placed at the opposing trap position in
SJWA is a managed freshwater wetlands located in the inland the circular trapline. Flow rates were checked at the beginning
southern California desert of Riverside County. This area is and end of each trapping period in order to validate a constant
known to have high mosquito activity and to contain diverse avian CO2 flow rate throughout the trapping period. Weather conditions
and mammalian fauna (Lura et al. 2012, CDFW 2016). Eight CDC were hot and dry. Temperatures recorded at the CIMIS (California
style suction traps (Model 512, JW Hock Inc., Gainesville, FL) Irrigation Management Information System) weather station near
were used to collect host-seeking Diptera. Six of the traps were Lake Perris averaged 35 (max) to 16 (min)° C during the sampling
augmented with CO2 from a 567 g cylinder tank equipped with periods (Figure 2). The prevailing wind direction was from the
a two-staged regulator, a flow restrictor to maintain a steady flow west and northwest with average daily wind speed of 8 kph.
of CO2, and vinyl plastic tubing placed so that CO2 was released For each collection night, insects captured in the two traps
just above the trap entrance. Traps were placed 30 m apart in a with the same CO2 flow rate were combined and sorted to species
circular trapline in an open field adjacent to mosquito production and sex, and enumerated, with the total number of females
and resting sites (Figure 1). This arrangement was used to ensure captured divided by two to give a mean per trap night count for
trap independence (Brown et al. 2008), unobstructed CO2 plume females of each species collected at each CO2 flow rate tested.
structure (Murlis and Jones 1981), and to minimize varying effects Mean counts were subsequently transformed to log10 (n+1) and
of vegetation on mosquito collections. CO2 concentrations (flow analyzed separately for low and high CO2 nights by ANOVA (Proc
rates) evaluated were 15, 47, 149, 473, and 1,495 ml of CO2 per GLM in SAS; Cary, North Carolina, U.S.A.) with week and CO2
min; these flow rates were half-log increases starting at 15 ml/ flow rate as independent variables. Means for significant variables
min, which was the lowest flow rate that could be maintained were separated using the post-hoc least significant difference
consistently by the regulators used in this study. (LSD) test. The use of traps with the mid-range CO2 flow rate
Trapping was conducted two nights per week during six weeks, (149 ml/min) on all nights allows for further comparison of mean
insect captures across all CO2 release rates.
To determine whether using traps with either low or high
CO2 flow rates in the trap line differentially impacted mid-range
flow rate trap captures, mean counts for mid-range flow rate
traps were paired by week (one low and one high CO2 night per
week) due to significant differences in mean counts among weeks
for some species, and evaluated using a paired t-test (SAS) for
differences among high and low CO2 nights. For species with mid-
range flow rate trap counts that did not differ between low and
high CO2 nights (all but An. hermsi Barr & Guptavanij which was
captured in very low numbers), the relationship of female capture
to CO2 concentration was subsequently determined across all CO2
release rates tested. To adjust for the natural variation in host-
seeking activity among weeks, it was necessary to transform each
mean trap count into a relative count that was a proportion of the
species-specific mean capture for the mid-range flow rate on the
same trap night.
Figure 1. Trap
CDC Suc'on Trapsarrangement at the wetlands of the San Jacinto To evaluate the effect of increasing CO2 concentration on
National Wildlife Area in southern California. female captures across both low and high CO2 concentrations
 19487134, 2017, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jvec.12238 by CAPES, Wiley Online Library on [03/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
46 Journal of Vector Ecology June 2017

80
Temp (Max) Temp (Min)
Humidity Precipita5on 0.92
70
0.82

60
0.72
Mean Rela6ve Humidity

50

Precipita6on (in)
0.62
Temp °C

40 0.52

0.42
30

0.32
20
0.22

10
0.12

0 0.02
13

13

/13

/13

/13

13

/13

/13

/13
7/1/

7/8/

8/5/
7/15

7/22

7/29

8/12

8/19

8/26
Figure 2. Meteorological data (Mean temperature, humidity and precipitation) taken from the CIMIS (California Irrigation Management
Information) weather station located near lake Perris.

tested, we corrected for the natural variation in host-seeking were Ae. vexans, An. hermsi, and An. franciscanus McCracken.
insect activity among nights by dividing the species-specific mean For all species, the number of females captured increased
capture for each CO2 flow rate on each trap night by the species- significantly with increasing CO2 concentration on both high
specific mean capture for the mid-range flow rate on the same CO2 nights [Cx. erythrothorax (F=230.64; df = 3,15; P<0
trap night to give a proportional mean per trap capture value for .0001), Cx. tarsalis (F=401.38; df=3,15; P<0.0001), Ae. vexans
each CO2 flow rate tested. Data was then analyzed by simple linear (F=17.99; df=3,15; P<0.0001), An. hermsi (F=12.16; df=3,15;
regression (R statistical package) against CO2 concentration. P=0.0003), An. franciscanus (F=14.42; df=3,15; P<0.0001), and
C. sonorensis (F=167.32; df=3,12; P<0.0001)] and low CO2 nights
RESULTS [Cx. erythrothorax (F=382.57; df=3,15; P<0 .0001), Cx. tarsalis
(F=195.83; df=3,15; P<0.0001), Ae. vexans (F=7.81; df=3,15;
A total of 34,196 female mosquitoes representing five species P=0.0023), An. hermsi (F=9.37; df=3,15; P=0.001), An. franciscanus
within three mosquito genera (Aedes, Anopheles, Culex) as well as (F=6.47; df=3,15; P=0.005), and C. sonorensis (F=11.59; df=3,12;
2,169 biting midges (C. sonorensis) were collected (Table 1). The P=0.0003)], although the increase in trap counts for Ae. vexans on
most abundant mosquitoes were Cx. tarsalis and Cx. erythrothorax low CO2 nights was only in comparison to the negative control (no
Dyar (57% and 40% of total mosquitoes collected, respectively). CO2) traps (Figure 3). While there was a trend for reduced mean
Other mosquito species collected in the order of their abundance counts in the mid-range CO2 flow rate traps on high CO2 nights
relative to low CO2 nights, this difference was not significant except
for An. hermsi (t=3.29; df=1,5; P<0.05), which was captured only
Table 1. Total number of biting flies captured in all traps. in very low numbers.
Species Females Males Blood fed There was a statistically significant linear relationship between
the CO2 concentration and the standardized proportional capture
Aedes vexans 816 0 1
for Cx. tarsalis (R2 = 0.93, p<0.001), Cx. erythrothorax (R2 = 0.88, p
Anopheles franciscanus 87 0 0 < 0.001), and C. sonorensis (R2 = 0.87, p < 0.001) with proportional
Anopheles hermsi 161 0 1 captures for species collected in lower numbers still predicted
Culex erythrothorax 13,796 68 6 primarily (R2=0.44-0.51) by the CO2 flow rate (Figure 4).
Simple linear regression of species-specific mean trap counts
Culex tarsalis 19,336 415 3
against CO2 flow rate showed that mosquitoes were captured at
Culicoides sonorensis 2,169 0 0 a similarly increasing rate over the range of CO2 concentrations
Total 36,365 483 11 tested (slope=0.0015-0.0031), while the more mammalophilic
 19487134, 2017, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jvec.12238 by CAPES, Wiley Online Library on [03/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Vol. 42, no. 1 Journal of Vector Ecology 47

500 A
A
Mean # Females per Trap

450 900
Cx. erythrothorax
400 800 Cx. tarsalis

Mean # Females per Trap

350 700
300 B 600
250
C 500 B
200 400
150
D
300 C
100 E 200 D
50 100 E
F F
0 0
0 15 47 149 473 1495 0 15 47 149 473 1495
CO2 flow rate (ml/min) CO2 flow rate (ml/min)

6 A
An. hermsi 4
5
A
Mean # Females per Trap

3.5 An. franciscanus

Mean # Females per Trap

4
A
3

2.5
3
C B 2
B
2 1.5 B
D 1 C
1
C
D 0.5
0 C
0
0 15 47 149 473 1495
0 15 47 149 473 1495
CO2 flow rate (ml/min)
CO2 flow rate (ml/min)

160 30
A A
Mean # Females per Trap

140 C. sonorensis 25
Mean # Females per Trap

120 A
20
100
15 B B
80 Ae. vexans
60 B 10
40 B
5
20 C
D D C
E 0
0
0 15 47 149 473 1495 0 15 47 149 473 1495
CO2 flow rate (ml/min) CO2 flow rate (ml/min)

Figure 3. Mean (SE) per trap capture of female mosquitoes and biting midges at CO2 flow rates tested. Bars with the same letter are not
significantly different (LSD test (P>0.05)).
 19487134, 2017, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jvec.12238 by CAPES, Wiley Online Library on [03/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
48 Journal of Vector Ecology June 2017

6 5
Propor<onal capture per trap

Propor,onal capture per trap

4.5 Cx. tarsalis
5 Cx. erythrothorax y = 0.0024x + 0.4408 4
R² = 0.8793
4
3.5
3
y = 0.0027x + 0.4328
3 2.5
R² = 0.9256
2
2 1.5
1
1
0.5

0 0
0 200 400 600 800 1000 1200 1400 1600 0 200 400 600 800 1000 1200 1400 1600

CO2 concentra<on CO2 concentra,on

5
10 An. franciscanus
Propor,onal capture per trap

Propor,onal capture per trap

4.5
9 An. hermsi
4 y = 0.0015x + 0.6928
8 R² = 0.51715
3.5
7 y = 0.0029x + 0.9454
3
6 R² = 0.44214
2.5
5
2
4
1.5
3
1
2
0.5
1
0
0
0 200 400 600 800 1000 1200 1400 1600
0 200 400 600 800 1000 1200 1400 1600

CO2 concentra,on CO2 concentra,on

9
Ae. vexans 35
Propor,onal capture per trap

8 C. sonorensis
Propor,onal capture per trap

y = 0.0025x + 1.0209
R² = 0.4936 30
7 y = 0.0146x + 0.1644
25 R² = 0.86638
6

5 20

4 15
3
10
2
5
1

0 0
0 200 400 600 800 1000 1200 1400 1600
0 200 400 600 800 1000 1200 1400 1600
CO2 concentra,on CO2 concentra,on
Figure 4. Scatter plots of standardized proportional counts per trap night (mean count per CO2 flow rate/mean count of the midrange CO2
flow rate on the same night) for female mosquitoes and biting midges captured. Counts at the midrange CO2 flow rate (149 ml/min) are
necessarily excluded from analysis as this counts were used to standardized counts for all other CO2 flow rates.
 19487134, 2017, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jvec.12238 by CAPES, Wiley Online Library on [03/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Vol. 42, no. 1 Journal of Vector Ecology 49

biting midge (C. sonorensis) was captured at a substantially greater McElligott (1989) reported a significant increase (six-fold)
rate with increasing CO2 concentration (slope=0.015). From the in Ae. vexans capture using ramp traps baited with 4,000 ml/
mid-range CO2 flow rate (149 ml/min) to the highest flow rate min in comparison with those baited with 1,000 ml/min. Our
tested (1,495 ml/min), there was a four-fold increase in capture observations for Cx. tarsalis are consistent with studies by Reeves
for all mosquito species, except An. hermsi which increased two to (1953) and Pfuntner et al. (1988) which reported significant
three-fold over this CO2 range, in contrast to a 25-fold increase in increases in capture of Cx. tarsalis when CO2 release rates were
the capture of C. sonorensis (Figure 4). increased from 250 to 1,000 ml/min, and with a study by Reisen et
A total of 483 males of two mosquito species, Cx. tarsalis al. (2000) which reported increased capture of Cx. tarsalis as CO2
(85%) and Cx. erythrothrorax (15%), was collected. Males of both release rates increased from 500 to 1,500 ml/min.
species were collected in similar numbers across all CO2 flow rates, It is important to note that the mosquito species collected
indicating males exhibit no increased or decreased orientation in the current study are all generalist feeders. Therefore, it is
toward traps based upon the presence or concentration of CO2 or perhaps not surprising to see that all the species sampled in
upon the differences in numbers of females collected during the this study responded to increasing CO2 concentrations with a
trapping period by traps with varying CO2 concentration (Figure similar rate of increasing capture. We had expected to capture Cx.
5). quinquefasciatus at this site, given the collection of this species
in an earlier study at this same location (Lura et al. 2012). It was
DISCUSSION anticipated that we might see a negative relationship between CO2
concentration and trap captures for this species given the earlier
Overall, CO2 concentration influenced the capture of all report by Reeves (1953) and by Mullens and Gerry (1998), and the
host-seeking Diptera sampled in this study, with increases in CO2 reduced feeding by this species on large mammals (horses) relative
concentration resulting in increasing capture of host-seeking to other mosquito species captured in this region of California
insects across all flow rates examined (15-1,495 ml/min). For all (Gerry et al. 2008).
mosquito species, the number of females captured increased at While it would be expected perhaps that a species like C.
an approximately similar rate with increasing CO2 concentration, sonorensis, which has a feeding preference for large mammals,
perhaps as a result of the similar opportunistic feeding behavior of would also have a strong positive relationship between number
these species. In contrast, captures of the highly mammalophilic of insects captured and CO2 concentration, it was nevertheless
biting midge (C. sonorensis) increased at a much faster rate, surprising how rapidly trap captures increased for C. sonorensis
particularly at the higher CO2 concentrations, probably reflecting when CO2 concentration increased from the mid-range flow rate
their preference for biting large mammals like cattle, sheep, deer, of 149 ml/min to the highest flow rate of 1,495 ml/min relative to
or horses as compared to the broader host range of the mosquito the more generalist feeding mosquitoes. The very large increase in
species captured. trap capture of C. sonorensis only at higher CO2 rates is consistent
This study is generally in agreement with earlier studies with previous studies by Mullens (1995) and Mullens and Gerry
that showed a positive relationship between numbers of insects (1998), and contrasts with the slower rate of increase at the highest
captured and CO2 concentration for blood feeding insect species CO2 flow rates for all the mosquito species captured in the current
that demonstrate a willingness to feed on a wide range of avian study. In this study, the increase in captures of C. sonorensis from
and mammalian hosts. Using traps baited with CO2 (0, 250, 400, the mid-range to highest CO2 rate was 25-fold relative to the
500, and 1,000 ml/min) Carestia and Savage (1967) found that increase in mosquito captures which were three-fold increases
increases in CO2 concentrations resulted in increasing capture over this same CO2 range.
of An. punctipennis, Ae. vexans, and Cx. salinarius. McIver and We hypothesize that higher CO2 release rates resulted in
an expanded odor plume covering a greater area compared to
20 lower release rates. This would increase the opportunity for more
18 Cx. erythrothorax
distant host-seeking mosquitoes to detect and initiate orientation
behaviors toward the CO2 source. More mammalophilic or
16
Mean # males collected

Cx. tarsalis
generalist species, like those captured in this study, might be
14 expected to continue to orient toward the odor source even at
12 the higher CO2 concentrations that would be encountered near
10 the trap. In contrast, ornithophilic species may be better adapted
per trap

8
to responding to lower CO2 release rates, more typical of small
animals and birds. These ornithophilic species might be expected
6 to be repelled by high concentrations of CO2 or just simply fail
4 to find the trap (CO2 source) under the high CO2 concentration
2 conditions near the trap. It would be interesting to repeat this study
0 at a site where mosquito species collected were not generalist host
0 15 47 149 473 1495 feeders but had a more narrow feeding preference for either birds
or large mammals.
CO2 flow rate (ml/min) This study highlights the importance of standardizing CO2
Figure 5. Mean [SE] per trap capture of male mosquitoes by CO2 release rates used with traps when conducting vector surveillance.
release rate. Medically important Diptera collected during this study showed
 19487134, 2017, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jvec.12238 by CAPES, Wiley Online Library on [03/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
50 Journal of Vector Ecology June 2017

significant differences in capture rate with increasing CO2 Gerry, A. C., T. M. Nawaey, P. B. Sanghrajka, J. Wisniewska,
flow rate. Variation in CO2 release rate might therefore impact and P. Hullinger. 2008. Hematophagous Diptera collected
mosquito surveillance interpretations regarding vector abundance from a horse and paired carbon dioxide-baited suction
and activity. For example, the mean trap count for Cx. tarsalis (an trap in southern California: Relevance to west Nile virus
important vector of WNV) would be expected to vary by 20% epizootiology. J. Med. Entomol. 45: 115-124.
or more across a range of probable CO2 flow rates when using Gillies, M. and T. Wilkes. 1974. The range of attraction of birds as
dry ice as the CO2 source (1,000 to 1,495 ml/min, according to baits for some West African mosquitoes (Diptera, Culicidae).
Mullens 1995). At our study site, this variation in CO2 flow rate Bull. Entomol. Res. 63: 573-581.
could have resulted in an increase of several hundred Cx. tarsalis Gillies, M. 1980. The role of carbon dioxide in host-finding by
captured per trap night between the low and high CO2 flow rate mosquitoes (Diptera: Culicidae): A review. Bull. Entomol.
within this range., The resulting increase in mosquito capture, due Res. 70: 525-532.
solely to variation in CO2 rate, might be incorrectly interpreted Lura, T., R. Cummings, R. Velten, K. De Collibus, T. Morgan, K.
as a worrisome increase in mosquito activity and pathogen Nguyen, and A. Gerry. 2012. Host (avian) biting preference of
transmission risk from one surveillance period to the next, when southern California Culex mosquitoes (Diptera: Culicidae). J.
the increase in mosquito capture might in fact be due solely to Med. Entomol. 49: 687-696.
variable CO2 release among trap nights. Mason, P. and J. Shemanchuk. 1990. Black Flies. Agriculture
While mosquito capture may vary with CO2 source (dry ice Canada Publication 1499/E, Communications Branch\
or compressed gas) (Reisen et al. 2000), we can reasonably expect Agriculture\Ottawa\K1A0C7: 1-56.
that the capture rate for the mosquito and biting midge species McIver, S. and P. McElligott. 1989. Effects of release rates on the
captured in this study would similarly increase with increasing CO2 range of attraction of carbon dioxide to some southwestern
flow rates, even if the actual rate of increase might differ somewhat Ontario mosquito species. J. Am. Mosq. Contr. Assoc. 5: 6-9.
among the CO2 sources. Additional field studies comparing Mokry, J. 1978. Progress towards the colonization of Cnephia
mosquito collections from dry ice baited traps with traps using mutate (Diptera: Simuliidae). Bull. Wld. Hlth. Org. 56: 455-
CO2 released from pressurized gas tanks are warranted to measure 456.
differences between these two methods of CO2 presentation. Mullens, B. 1995. Flight activity and response to carbon dioxide of
The typical CO2 output of a dry ice baited suction trap falls Culicoides variipennis sonorensis (Diptera: Ceratopogonidae)
within the range of flow rates tested in this study (Mullens 1995, in southern California. J. Med. Entomol. 32: 310-315.
Eshun et al. 2016). If dry ice is utilized with the trap, the CO2 Mullens, B. and A. Gerry. 1998. Comparison of bait cattle and
release rate could be estimated by the loss in dry ice weight due carbon dioxide-baited suction traps for collecting Culicoides
to sublimation of the gas during the trapping period. Even if CO2 variipennis sonorensis (Diptera: Ceratopogonidae) and Culex
rates vary somewhat by trap night, collections could be adjusted quinquefasciatus (Diptera: Culicidae). J. Med. Entomol. 35:
based upon the direct relationship between CO2 concentrations 245-250.
and mosquito capture rates. Our study suggests that if standard Murlis, J. and C. Jones. 1981. Fine-scale structure of odour plumes
trap designs and suitable collection sites are utilized to minimize in relation to insect orientation to distant pheromone and
design and site effects (Resien et. al. 2000), the host seeking activity other attractant sources. Physiol. Entomol. 6: 71-86.
of species captured in this study could be effectively modeled, Pfuntner, A., W.K. Reisen, and M.S. Dhillon. 1988. Vertical
even with variation in CO2 release, so that surveillance data could distribution and response of Culex mosquitoes to differing
be more reliably used. concentrations of carbon dioxide. Proc. Calif. Mosq. Vector
Contr. Assoc. 56: 69-74.
REFERENCES CITED Reeves, W.C. 1953. Quantitative field studies on a carbon dioxide
chemotropism of mosquitoes. Am. J. Trop. Med. Hyg. 2: 325-
Brown, H.E., M. Paladini, R.A. Cook, D. Kline, D. Barnard, and 331.
D. Fish. 2008. Effectiveness of mosquito traps in measuring Reisen, W.K., R.P. Meyer, R.F. Cummings, and O. Delgado.
species abundance and composition. J. Med. Entomol. 45: 2000. Effects of trap design and CO2 presentation on the
517-521. measurement of adult mosquito abundance using Centers
Carestia, R. and L. Savage. 1967. Effectiveness of carbon dioxide for Disease Control-style miniature light traps. J. Am. Mosq.
as a mosquito attractant in the CDC miniature light trap. Contr. Assoc. 16: 13-18.
Mosquito News 27: 90-92. Roberts, R. 1972. Relative attractiveness of CO2 and steer to
Eshun, O., A. Gerry, and W.K. Hayes. 2016. Mosquito capture rate Tabanidae, Culicidae, and Stomoxys calcitrans (L.). Mosq.
using CO2-baited traps in relation to distance from water and News 32: 208-211.
height: implication for avian disease transmission. J. Med. Rohe, D.L. and R.P. Fall. 1979. A miniature battery-powered CO2
Entomol. http://dx.doi.org/10.1093/jme/tjw120. baited light trap for mosquito-borne encephalitis surveillance.
Frommer, R., B. Shiefer, and R. Vavra,. 1976. Comparative effects Bull. Soc. Vector Ecol. 4: 24-27.
of CO2 flow rates using modified CDC light traps on trapping Sudia, W.D. and R.W. Chamberlain. 1962. Battery-operated light
adult black flies (Simuliidae: Diptera). Mosq. News 36: 355- trap, an improved model. Mosq. News 22: 126-129.
358.