Predation by the Introduced Phoretic Mite~

Macrocheles peregrinus (Acari: Macrochelidae)~ on the

Buffalo Fly~ Haematobia irritans exigua
(Diptera: Muscidae)~ in Australia 1

J. P. ROTH," A. MACQUEEN," AND D. E. BAY'

Environ. Entomol. 17(3): 603-607 (1988)

ABSTRACT The effectiveness of Macrocheles peregrinus (Krantz) as a predator of buffalo
fly, Haematobia irritans exigua (De Meijere), was examined. In laboratory tests, mites were
confined with eggs or larvae of buffalo fly to assess the effect of prey location and predator I
prey ratio on fly survival, to determine the main life stage attacked, and to determine the
prey preference of the mite in the presence of other fly species. In field tests, impact of the
mite on buffalo fly survival was measured in natural field pats enclosed in screen cages.
Buffalo fly mortality was highest when eggs of this species were placed on top of dung pats
rather than beneath them. Maximum mortality obtained was 83%; there was no increase in
mortality at predator Iprey ratios above one mite per two fly eggs. The mite caused highest
mortality when exposed to hatching eggs and young larvae of buffalo fly, but it showed a
strong preference for eggs of other dipteran species that have softer egg chorions. Mites
acting alone caused an average of 33% suppression of buffalo fly breeding in field pats.
Macrocheles peregrinus was judged to be a relatively ineffective predator of buffalo fly.

KEY WORDS Arachnida, phoretic predacious mite, Haematobia irritans exigua, buffalo fly

IN 1975, THE CSIHO Division of Entomology began 1983, Doube et a!. 1986). Subsequent to releases of
a study of predacious, phoretic mites associated M. peregrinus at Rockhampton, Queensland in
with native and established dung beetles in Aus- December 1980 and again in February 1981 near
tralia. The purpose of these investigations was to the Adelaide River, Northern Territory, Wallace
enhance biological control of the bush fly, Musca & Holm (1983) found that this mite was established
vetustissima (Walker), and the buffalo fly, Hae- and widely distributed in central Queensland and
matoma irritans exigua (De Meijere) (Wallace et the northeastern portion of the Northern Territory.
a!. 1979). The control of these two pests, along with The mites were primarily associated with the in-
pasture improvement, was the primary goal of the troduced dung beetle Euoniticellus intermedius
Australian dung beetle project (Waterhouse 1974). (Reiche), but they were also recovered from several
Wallace et al. (1979) found that Macrocheles gla- other introduced scarabs and one native scarab
ber (Muller), a mite abundant on native dung bee- species.
tles in southeastern Australia, is an effective pred- Our report describes studies designed to deter-
ator of the bush fly. Neither M. glaber nor any mine the effectiveness of M. peregrinus as a pred-
similar mite was found to be present within the ator of the buffalo fly in central Queensland.
range of the buffalo fly in Queensland and the
Northern Territory.
Krantz (1981) described Macrocheles peregri- Materials and Methods
nus (Krantz) from collections made in South Af- A laboratory colony of M. peregrinus was es-
rica, Namibia, and Kenya. This mite was selected tablished at Rockhampton in 1984 from local field
for introduction to Australia because of its wide collections. Laboratory colonies of the mites were
geographic distribution, its relative abundance in maintained on fresh cattle dung shaped into 200-g
Africa, and its range of phoretic hosts that included pats and placed on 25 cm of clean sifted sand in
most of the African dung beetles that have been a plastic container (30 by 20 by 12 cm). The sides
established in northern Australia (Wallace & Holm of the container were coated with fluon (ICI, Syd-
ney, Australia) to prevent mites from escaping. Col-
onies were held at 26°C and a 14:10 (L:D) photo-
I This artic!p rpports the resultsof research only. Mention of a period. Mites were supplied with bush fly eggs from
proprit'lary product dol'Snot constitute an endorsement or a rec-
onlllU'oJation for its use by USDA. a laboratory colony four times a week. Fresh dung
'Veterinary Toxicologyand EntomologyResearchLaboratory, pats were added to the colonies twice a week. The
CSDA~ARS,P.O. Drawer GE, College Station, Tex. 77841. new dung pats were placed with their edges touch-
3 Divisionof Entomology,CSIHO,Box5545, RockhamptonMail
ing the older pats to facilitate translocation by the
Centrt', Quepnsland 4702, Australia.
• Department of Entomology, Texas A&M University,College mites. Old desiccated dung pats were removed as
Station, Tex. 779·tQ. needed. All the mites used in experiments were
604 ENVIRONMENTAL ENTOMOLOGY Vol. 17, no. 3

adult females; male M. peregrinus have a com- mites were exposed only to the eggs; treatment 3,
paratively short life span, and the sex ratio heavily in which 20 mites were exposed to hatching eggs
favors the females. Mites to be used in tests were and first-instar larvae; and treatment 4, in which
removed from the colony and held without feeding 20 mites were exposed to hatched larvae that had
overnight on dung that had been frozen to destroy been allowed 2 h to burrow into the dung. In treat-
any food source. ment 2, the mites and eggs were placed on the
H. irritans exigua normally oviposits beneath dung at the same time. At approximately I h before
cow pats immediately after defecation. Oviposition hatch (as indicated by the first appearance of larval
occurs not only on the dung but on soil and ground striations in the eggs), the eggs were carefully re-
litter that are in contact with the dung. Natural moved and transferred to the surface of dung pats
oviposition was difficult to obtain for laboratory in containers that had no mites. In treatment 3,
tests; therefore, dung pats were seeded with known mites were not added to the containers until larval
numbers of Haematobia eggs using a standard striations appeared within the eggs; thus hatching
technique that simulated natural oviposition. This eggs and larvae were exposed. In treatment 4, mites
was done by forming 150 g of fresh bovine dung were added to the containers 2 h after hatch.
into cylindrical pats and supporting each pat on a Prey Preference Tests. The first prey prefer-
bamboo grid placed on 5 cm of dry sand in a I-liter ence test was conducted by exposing eggs of buffalo
container. The grids were constructed from bam- fly and Orthellia lauta (Wiedemann) to adult fe-
boo skewers (2 mm diameter by 8 cm long) glued male M. peregrinus on moistened filter paper in a
together by epoxy cement. Each grid consisted of sealed Petri dish. After 24 h, the number of hatched
a lower row of four parallel bamboo sections spaced eggs and live larvae of each prey species was count-
1.5 cm apart, with a similar upper row glued at a ed. Treatments included a buffalo fly control with
slight angle. Egg seeding was accomplished by im- 20 eggs per Petri dish, an O. lauta control with 20
printing a grid on the surface of each test pat and eggs per Petri dish, a treatment with 20 buffalo fly
placing a known number of eggs on the dung at eggs and four mites per Petri dish, and a treatment
approximately evenly distributed intervals along with 20 eggs of each fly species and eight mites
the lines of contact with the grid. The dung pat per Petri dish. Each treatment in this test was rep-
and grid were then inverted and placed egg-side licated five times.
down on the sand in the test containers. The top Subsequent preference tests with buffalo fly and
inner sides of each test container were coated with bush fly were conducted using the previously de-
fluon to prevent mites from escaping. scribed technique for seeding buffalo fly eggs under
Buffalo fly eggs were obtained by netting adult the dung pat. In treatments that contained bush
flies from cattle and inducing oviposition by CO2 fly eggs, small indentations were made around the
anesthesia as described by Macqueen et al. (1986). lower edges of the dung pats near the point of
Prey Location. A test was designed to determine contact with the soil. Bush fly eggs from a labo-
the effect of prey location on the ability of M. ratory colony were placed in clusters in these in-
peregrinus to prey on buffalo fly eggs. In this test, dentations to simulate natural oviposition sites. Two
half the pats were prepared as described for the test designs were used: the treatments in the first
standard technique with 30 buffalo fly eggs on the test design consisted of a buffalo fly control with
underside, and the other half were prepared iden- 50 eggs per pat, a bush fly control with 50 eggs per
tically except that the buffalo fly eggs were placed pat, a combined control with 25 eggs of each species
randomly on the top of the dung. The pats then per pat, 50 buffalo fly eggs with 20 mites per pat,
were infested with either 0, IS, or 45 M. peregri- 50 bush fly eggs with 20 mites per pat, and 25
nus, respectively, per pat. After 6 d, Haematobia buffalo fly eggs with 25 bush fly eggs and 20 mites
pupae were collected by sifting the sand and ex- per pat. In this test, therefore, the total prey jpred-
amined the dung in each test container. ator ratio was constant. The treatments in the second
Predator/Prey Ratios. Dung pats and con- test design were similar, except that 40 eggs of each
tainers were prepared by the previously described prey species per pat were used. Thus, the combined
method for simulating natural oviposition. The prey-control and the combined prey-mite treat-
number of buffalo fly eggs per test container was ment had a total of 80 eggs per pat. Treatments
held constant at 30 eggs, and the number of mites with mites had 15 mites per pat in this test.
was varied to change predator jprey ratios. The With the exception of the preliminary test on
tests included a control with no mites and eight prey preference (which had five replications), all
mite treatments ranging from 3 to 60 mites per the laboratory tests were replicated six times. Anal-
container. ysis of variance was conducted for a completely
Buffalo Fly Stage Attacked. A test was con- random design, and treatment means were sepa-
ducted to determine the period of buffalo fly sus- rated using Duncan's multiple range test (Duncan
ceptibility to predation by M. peregrinus. Test con- 1955).
tainers were prepared with dung as before, except Survival in the Field. A field trial of the impact
that buffalo fly eggs (30) were placed on top of the of M. peregrinus on buffalo fly was conducted in
dung pats. Treatments included a control (I) that 1985. This trial consisted of four tests which were
had buffalo fly eggs only; treatment 2, in which 20 conducted on 13 February, 10 April, 1 May, and
J line 1988 ROTH ET AL.: M. peregrinus PREDATION ON BUFFALO FLY 605

TobIe 1. E£reets of egg loeotion on predotion of M. Table 2. Buffalo fly survival as a function of exposure
peregrinus on the buffalo fly to varying numbers of M. peregrinus in the laboratory

No. of % buffalo fly No. of No. of Mites per

Ep.p.location on pat % survival
M. peregrinus survivala mites prey egg
Top o 41. 7a 0 30 0 73.9 ± 7.4
Bottom o 42.2a 3 30 0.1 37.3 ± 8.7
Top 15 l.ld 6 30 0.2 27.3 ± 14.8
Bottom 15 30.ob 9 30 0.3 25.3 ± 12.6
Top 45 1.7d 12 30 0.4 24.0 ± 5.5
Bottom 45 18.3c 15 30 0.5 22.2 ± 13.9
30 30 1.0 12.2 ± 7.8
a M,'ans not followed by the same letter are significantly dif- 45 30 1.5 12.8 ± 8.9
[,'n'nt (P = 0.05 level, Duncan's [1955] multiple range test). 60 30 2.0 20.5 ± 12.4

15 May. The following procedure was used in each

test. Cattle dung pats were marked immediately [auta and for the bush fly over buffalo fly. Neither
upon defecation. The pats were observed until ovi- of these species habitually oviposit directly under
positing buffalo flies left the dung (ca. 3 min). After the dung.
oviposition occurred, the pats were covered with The results of the predator /prey ratio tests (Ta-
mesh-covered metal cones to exclude all dung fau- ble 2) indicate that all the mite treatments caused
na (Macqueen & Beirne 1975). Ten pats were cov- significantly more mortality than occurred in the
ered on each test date. Five of these pats were control, but even the highest mite/prey ratios did
seeded with 500 M. peregrinus from the laboratory not completely suppress buffalo fly development to
colony before being covered; these mites had been the degree that predation on top-seeded eggs did
held overnight without feeding. The remaining five in the egg placement test. This would seem to in-
pats served as the controls. The test pats and a dicate that when eggs were seeded underneath the
shallow divot of the soil beneath the pats were dung pat, some of the buffalo fly eggs were not
collected after 48 h in the pasture and held in an exposed to mite predation. Settling of the dung
insectary for 3 wk in cages fitted with emergence could make some of the buffalo fly eggs and larvae
traps to collect emergent Diptera. The data were inaccessible to the mites. To what degree this might
analyzed by ranking the number of buffalo flies occur with natural oviposition is not known.
collected from each control pat in descending order Results of the stage preference test (Table 3)
from highest to lowest and performing a paired t indicated that the only treatment in which there
test with similarly ranked mite treatments. Paired was a significant reduction in the number of buffalo
t tests were used to separate treatment means for fly pupae collected was that in which mites were
each test and for combined treatment means for added just prior to egg hatch. Exposure of the buf-
the whole trial. The fauna-exclusion cones used in falo fly eggs to the mites during the pre hatch period
the field trial were intended to keep out all dung did not cause a significant reduction in buffalo fly
fauna except buffalo flies. However, several species survival. Apparently there was some predation
of Diptera (such as Orthellia [auta and Sepsidae when mites were added 2 h after hatch, but this
spp.) were able to colonize the pats during the time was not statistically significant. These results differ
when they were still exposed to buffalo fly ovipo- from those of Doube et al. (1986), who reported
sition. The Sepsidae spp. were collected so consis- that predation on buffalo fly eggs did occur when
tently that we decided to analyze the data on sepsid M. peregrinus was exposed to eggs seeded on top
emergence as well as that on buffalo fly. of 50-ml dung pats. The high level of predation
that occurred on the top-seeded treatment in our
prey location test (Table 1) was similar to the level
Results and Discussion
of predation that occurred on bush fly eggs in the
The results of the buffalo fly egg placement test prey preference test (see Table 5). It is unlikely
(Table 1) indicate that prey location on the dung
pat has a direct effect on its susceptibility to pre-
dation of M. peregrinus. Survival was less than 2%
Table 3. Period of immature buffalo fly susceptibility
in both treatments in which eggs on the surface of to predation by M. peregrinus
the dung were exposed to the mites. In comparison,
survival in the mite treatments with bottom-seeded Exposure period
No. of buffalo
% survival
eggs was 30% with 15 mites per pat and 18% with fly pupaea
45 mites per pat. Survival in the two control treat- None (control) 11.7a 39.8
ments was about 42% for both top- and bottom- 16 h prehatch 1O.3a 34.4
2 h prehatch to pupae 4.8b 16.1
seeded pats. The mites could be more adapted to
2 h posthatch to pupae 8.2a 27.2
predation on species which do not oviposit under-
neath the dung pat. In fact, in our prey preference a Means not followed by the same leller are significantly dif-
tests, M. peregrinus showed a preference for O. ferent (P = 0.05; Duncan's [1955] multiple range test).
606 ENVIRONMENTAL ENTOMOLOGY Vol. 17, no. 3

Table 4. Prey preference by M. peregrinus for eggs tion in buffalo fly survival by the mites, but this
and larvae of buffalo fly and O. lauta exposed to Ihe miles may have been masked by low buffalo fly survival
on moistened filler paper
in the 'control treatments. In contrast, bush fly sur-
O. Lauta vival was drastically reduced by the mites.
Buffalo fly There was better survival in the buffalo fly con-
%
Treatments % % %
live
trol in the test in which the total number of prey
egg live egg was variable. In this test the numbers of prey species
lar-
hatch larvae hatch
vae were high enough so that there was some compe-
20 buffalo fly eggs (control) 89a 87.5a tition between buffalo fly and bush fly larvae. Bush
20 O. Lauta eggs (control) 81a 0 fly survival in the combined-fly control was signif-
20 buffalo fly eggs + 4 mites 86a 46.7c icantly lower (60.4%) than in the bush fly control
20 buffalo fly eggs + 20 O. (81.6%). Buffalo fly survival in the combined con-
Lauta eggs + 8 mites 95a 68.5b 53.6b 0
trol was not significantly different from survival in
Means in columns not followed by the same letter are signifi- the buffalo fly-only control. In the buffalo fly-mite
cantly different (P = 0.05, Duncan's [1955] multiple range test). combination, mites caused a significant reduction
in buffalo fly survival (28.7%) compared with buf-
falo fly controls (64.6 and 57.9%) and the combined
that such high mortality could be accounted for by prey and mite treatment (46.6%). As in the pre-
predation on the buffalo fly larvae alone; therefore, ceding test, bush fly survival in the presence of
it is probable that under some circumstances the mites (1.6 and 2.1%) was significantly lower than
mites will prey on prehatched buffalo fly eggs seed- bush fly or buffalo fly survival in any other treat-
ed on the top of dung pats. ment. The significantly higher survival of buffalo
Results of the initial prey preference test are flies in the combined prey-plus-mite treatment
presented in Table 4. The presence of the mites (46.6% compared with 28.7%) indicates that the
had no significant effect on the mean percentage presence of a preferred alternative prey species can
of egg hatch of buffalo fly eggs. The mean per- reduce M. peregrinus predation on the buffalo fly.
centage of the buffalo fly larvae alive at 24 h was The results of both tests demonstrate that M. per-
significantly lower than the control in the treatment egrinus is a much more effective predator of bush
with buffalo fly, O. [auta, and mites. The presence fly than buffalo fly. Although survival in the control
of O. [auta caused a significant reduction in the treatments in these laboratory tests was low, this is
degree to which buffalo fly larvae were preyed a typical result when Haematobia spp. eggs are
upon in the combined-prey treatment. The mean artificially seeded in dung pats; variability is not
percentage of egg hatch of O. [auta eggs was sig- likely to be improved by additional replication.
nificantly reduced in the combined treatment com- Thomas & Morgan (1972) found similar levels of
pared with the control. 0., [auta larvae did not mortality in control treatments with the horn fly,
survive long on the moistened filter paper, and no Haematobia irritans (L.). In the present series of
live larvae were found after 24 h in either the tests, there were a total of 36 control dung pats,
control or the combined prey treatment. and survival ranged from 31.4 to 73.9% with a
The results of the two tests for preference of M. mean of 49 ± 16.4%.
peregrinus for bush fly or buffalo fly are presented The results of the field trial (Table 6) indicate
in Table 5. In the test in which the number of total that there was considerable variation in the number
prey was constant, there was no significant reduc- of buffalo fly eggs deposited within and between

Table 5. Prey preference by M. peregrinus for buffalo fly and bush fly in two laboratory tests that simulated natural
oviposition by these prey species

No. total prey constanta No. total prey variableb

Buffalo fly Bush fly Buffalo fly Bush fly
Treatments
% survival % survival % survival % survival
(no. eggs) (no. eggs) (no. eggs) (no. eggs)
I-Buffalo fly control 31.6a (50) 64.6b (40)
2-Bush fly control 7l.6a (50) 81.6a (40)
3-Buffalo fly + bush fly 32.0a (25) 66.0a (25) 57.9b (40) 60.4b (40)
4-Buffalo lIy + mites 25.3a (50) 28.7d (40)
5-Bush lIy + mites 4.3c (50) 1.6e (40)
6-Buffalo lIy + bush lIy + mites 25.3a (25) Oc (25) 46.6c (40) 2.1e (40)

Means are not compared between the two tests. Within the tests, means not followed by the same letter are significantly different
(P = 0.05. Duncan's [1955] multiple range test).
a No. prey (eggs). 50. This consisted of either 50 buffalo fly eggs and 50 bush lIy eggs. or 25 buffalo lIy eggs and 25 bush lIy eggs.
The mite treatments received 20 mites per pat.
b No. prey (eggs). 40 for both the buffalo lly and the bush lly, regardless of whether the treatment used one or both species. Thus,
combined species treatments had a total of 80 prey eggs, and the single species treatments had a total of 40 prey eggs. The mite
treatments received 15 mites per pat.
June 1988 ROTH ET AL.: M. peregrinus PREDATION ON BUFFALO FLY 607

Table 6. Effects of predation by M. peregrina. on H. irriran. exigua

Mean no. of adult flies"

!Jatl' H. irritans exigua Sepsidae spp.
Control Mitesb % suppression Control Mitesb % suppression

.13 Fl'b. .19H5 340.6 330.6 2.9 77.2 69.4 .10..1

10 Apr . .19H5 .188.6 77.8 58.7 22.1 48.6 78
t ~Iay .19H5 .126.2 59.4 52.9 74 .19.4 73.8
15 ~Iay .19H5 60.4 .10.4 82.8 .128.6 9.4 92.7*
TotalC 3,577 2,39.1 2,454 734
(h,'ralllO,'anc 179.8 119 33.2* 123 37 70.1*

" Vallit's arl' me'ans of fivt' dung pats.

h Eae'h dun~ pat s<,<,de'dwith 500 At. peregrinas from the laboratory colony .
•.Total and o\'<,rall m<,an are from combined data on 20 dung pats per treatment; *, significant (P = 0.05, paIred t test).

tests. The results of the 13 February test in partic- grinus may not be a very effective predator of the
ular were influenced by the emergence of 1,500 buffalo fly. Natural prey location, prey stage at-
buffalo flies from one of the mite-treated pats. tacked, preference for alternative prey species, and
However, the fact that the mites also performed the need for high predator /prey ratios tend to limit
poorly against Sepsidae in the 13 February test, predation by these mites on this species. The es-
could indicate that these mites were in poor con- tablishment and dispersal of M. peregrinus is more
dition or that adverse climatic factors could have likely to have an effect on the bush fly (for which
been involved. Only in the 15 May test was suppres- the mites have demonstrated a preference), which
sion of the buffalo flies statistically significant; over- oviposits in more accessible parts of the dung pat.
all, mites caused a significant reduction of 33% (P =
0.05) in buffalo fly emergence. This result is prob-
References Cited
ably not representative because of the lack of
suppression in the 13 February test. If the results Doube, B. M., A. Macqueen & K. A. Huxham. 1986.
of just the 10 April, 1 May, and 15 May tests are Aspectsof the predatory activity of Macrocheles per-
averaged, the resulting predation rate would be egrinus (Acarina: Macrochelidae) on two species of
64.8%. Doube et al. (1986) reported 66 and 54% Haematobia fly(Diptera: Muscidae),pp. 132-141. In
suppression, respectively, in two field tests. Entomological Society of America Miscellaneous
Publication 61.
Although the results of the 10 April and I May Duncan, D. B. 1955. Multiple range and multiple F
tests indicate substantial predation, mites signifi- tests. Biometrics 11: 1-42.
cantly reduced sepsid emergence only in the IS Krantz, G. W. 1981. Two new Claber group species
May test. The pooled data, including the 13 Feb- of Macrocheles (Acari: Macrochelidae) from southern
ruary test, indicated that the mites caused a 70% Africa. Int. J. Acarol. 7: 3-16.
reduction in sepsid emergence. The overall number Macqueen, A. & B. P. Beirne. 1975. Influence of
of sepsids emerging from control pats was 30% other insects on production of horn fly, Haematobia
lower than the number of buffalo flies from these irritans (Diptera: Muscidae), from cattle dung in
south-central British Columbia. Can. Entomol. 107:
pats. Therefore, M. peregrinus preyed more heavi-
1255-1264.
lyon the less numerous Sepsidae than on the buffalo Macqueen, D. F., M. M. H. Wallace & B. M. Doube.
flies. 1986. Seasonalchanges in favorability of cattle dung
The presence of alternate prey must be taken in central Queensland for three species of dung-
into account in relating the results of the field trial breeding insects. J. Aust. Entomol. Soc. 25: 23-29.
to the potential impact of M. peregrinus on buffalo Thomas, G. D. & C. E. Morgan. 1972. Field-mortality
fly populations. In these tests, Sepsidae and other studies of the immature stages of the horn fly in
Diptera had only a limited time to colonize the Missouri. Environ. Entomol. 1: 453-459.
pats before the mesh cones were applied. Under Wallace, M. M. H. & E. Holm. 1983. Establishment
and dispersal of the introduced predatory mite, Mac-
normal field conditions, there would be larger
rocheles peregrinus Krantz, in Australia. J. Aust.
numbers of alternate prey available; this could re- Entomol. Soc. 22: 345-348.
duce the level of predation on the buffalo fly. A Wallace, M. M. H., M. Tyrdale-Bisco & E. Holm. 1979.
second factor that must be considered is that the The influence of Macrocheles glaber on the breeding
numbers of mites per pat used in these tests rep- of the Australian bushfly, Musca vetustissima, in cow
resent a level of colonization that may occur in- dung, pp. 217-222. In Recent advances in acarology,
frequently in the field, Wallace & Holm (1983) vol. 2.
reported that weekly mite collections in pitfall traps Waterhouse, D. F. 1974. The biological control of
exceeded 500 per trap on only 5 of 29 occasions. dung. Sci. Am. 230(4): 101-108.
We assume that similar numbers would colonize
Received for publication 31 July 1987; accepted 11
dung pats. February 1988.
The results of our study indicate that M. pere-