Cold Susceptibility and Disinfestation of Bactrocera invadens

(Diptera: Tephritidae) in Oranges

Author(s): T. G. Grout, J. H. Daneel, S. A. Mohamed, S. Ekesi, P. W. Nderitu, P. R.
Stephen, and V. Hattingh
Source: Journal of Economic Entomology, 104(4):1180-1188. 2011.
Published By: Entomological Society of America
DOI: 10.1603/EC10435
URL: http://www.bioone.org/doi/full/10.1603/EC10435

BioOne (www.bioone.org) is a a nonprofit, online aggregation of core research in the

biological, ecological, and environmental sciences. BioOne provides a sustainable online
platform for over 170 journals and books published by nonprofit societies, associations,
museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates
your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and non-commercial
use. Commercial inquiries or rights and permissions requests should be directed to the
individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers,
academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.
 COMMODITY TREATMENT AND QUARANTINE ENTOMOLOGY

Cold Susceptibility and Disinfestation of Bactrocera invadens

(Diptera: Tephritidae) in Oranges
T. G. GROUT,1,2 J. H. DANEEL,1 S. A. MOHAMED,3 S. EKESI,3 P. W. NDERITU,3
P. R. STEPHEN,1 AND V. HATTINGH4

J. Econ. Entomol. 104(4): 1180Ð1188 (2011); DOI: 10.1603/EC10435

ABSTRACT To develop a cold disinfestation treatment for the fruit ßy Bactrocera invadens Drew,
Tsuruta & White (Diptera: Tephritidae) that is rapidly spreading across Africa, research was con-
ducted in Nairobi, Kenya, using ßies from a laboratory culture and ÔValenciaÕ orange (Citrus sinensis
L. Osbeck) as the host. The developmental rate of B. invadens in Valencia oranges was determined
at 28⬚C, and the third instar was found to be the least susceptible of the egg and larval life stages to
cold treatment at 1.1⬚C in oranges. When 22,449 B. invadens third instars were exposed in oranges to
a cold treatment with an approximate midpoint of 1.1 ⫾ 0.5⬚C, the results suggested that a period of
16 d would be worthwhile verifying on a larger scale in oranges. Results from the Þrst replicate of 16,617
larvae showed no survivors, but the second replicate of 23,536 larvae had three survivors. Because a
longer cold treatment based on a mean temperature of 1.1⬚C would create logistical difÞculties for
some export markets, further replicates were conducted at an approximate midpoint of 0.5⬚C and at
mean hourly maximum of 0.9 ⫾ 0.5⬚C, for 16 d. After three replicates, in which 65,752 B. invadens third
instars in total were treated with no survivors, the Japanese requirement of 99.99% mortality at the
95% conÞdence level was surpassed. The following treatment protocol for B. invadens larvae in oranges
can therefore be recommended: fruit pulp to be maintained at temperatures of 0.9⬚C or lower for 16
consecutive days.

KEY WORDS Bactrocera invadens, cold disinfestation, quarantine, commodity treatment, phyto-
sanitary

Bactrocera invadens Drew, Tsuruta & White (Diptera:
Tephritidae) (Drew et al. 2005) was Þrst detected on
the African continent in 2003 (Lux et al. 2003) at the
Kenyan coast by using a protein-baited trap and later
described in 2005 as a pest that was completely new to
science (Drew et al. 2005). Since then, established
populations have been found as far west as Senegal
and Cape Verde, to the north as far as Sudan, to the
east as far as the Comoro Islands and south to Namibia
and Mozambique (⬇18⬚ S) (Ekesi and Muchugu 2007;
Rwomushana et al. 2008b). This ßyÕs primary host
seems to be mango, Mangifera indica L. (Anacardi-
aceae) (Rwomushana et al. 2008a,b), in which it is
outcompeting the native fruit ßy Ceratitis cosyra
(Walker) in tropical Africa at low altitudes where the
relative humidity is high (Ekesi et al. 2006, 2009; Ndi-
aye et al. 2008). B. invadens also has been recorded as
a pest of various Citrus species (Rwomushana et al.
2008b), but in Tanzania, grapefruit, Citrus paradisi
1 Citrus Research International, P.O. Box 28, Nelspruit 1200, South
Africa.
2 Corresponding author, e-mail: tg@cri.co.za.
3 International Centre of Insect Physiology and Ecology, P.O. Box
30772Ð 00100, Nairobi, Kenya.
4 Citrus Research International, Department of Conservation Ecol-
ogy and Entomology, Stellenbosch University, P.O. Box 2201, Matie-
land 7602, South Africa.
Macfad., is more likely to be infested than other Citrus
species (Mwatawala et al. 2006).
It is inevitable that B. invadens will eventually in-
vade South Africa because it has recently been de-
tected along the borders with Zimbabwe and Bo-
tswana (IPPC 2010). If it becomes established, export
fruit markets will require assurance that there will be
no live fruit ßy in the exported fruit, or if there are, that
they will not survive in the country that is receiving
the fruit. The following research was therefore con-
ducted at the International Centre of Insect Physiol-
ogy and Ecology (ICIPE) in Nairobi, Kenya, to de-
termine whether the 16-d cold treatment at a mean of
1⬚C for the Mediterranean fruit ßy, Ceratitis capitata
(Wiedemann), that was previously veriÞed by Grout
et al. (2011), or the USDAÐAPHIS postharvest cold
treatment schedule T107-a (APHIS 2010) requiring
14 d at or below 1.11⬚C for C. capitata, would give
adequate control of B. invadens larvae in ÔValenciaÕ
orange (Citrus sinensis L. Osbeck).
Earlier research in Kenya has led to the develop-
ment of a suitable artiÞcial diet (Ekesi et al. 2007) and
provided information on the development and opti-
mal rearing temperature for B. invadens, found to be
28⬚C (Rwomushana et al. 2008b). This allowed for the
provision of suitable numbers of B. invadens larvae for
this postharvest research with oranges. Although very

0022-0493/11/1180Ð1188$04.00/0 䉷 2011 Entomological Society of America

August 2011 GROUT ET AL.: B. invadens COLD SUSCEPTIBILITY IN ORANGES
little citrus is exported from Kenya as fresh fruit, South
Africa is the second largest exporter of citrus in the
world after Spain (FAO 2006). Some South African
fruit is exported to countries that require either a cold
treatment for the moth Thaumatotibia leucotreta
(Meyrick) or for the Mediterranean fruit ßy. The
United States of America requires a cold treatment for
the former pest of either 22 or 24 d (depending on the
time of year) at or below ⫺0.55⬚C (Schedule T107-k,
APHIS 2010), and some other countries insist on the
same requirement. This extreme treatment will con-
trol any fruit ßy so the need for citrus exports to the
United States having to meet the less stringent re-
quirements for fruit ßy is largely irrelevant. The ob-
jective of our study was therefore to primarily develop
a treatment that would satisfy the Japanese require-
ment of 99.99% mortality at the 95% conÞdence level.
However, we decided to aim above this requirement
and obtain 99.9968% mortality (Probit 9), although not
at the 95% conÞdence level that would have required
the treatment of 93,613 larvae without any survivors
(Follett and Neven 2006).
Materials and Methods
This research was conducted in four phases. The
Þrst phase was to determine the larval development
rate in oranges at an ideal rearing temperature. The
second phase was to determine which developmental
stage (egg or larval) would be the least susceptible to
cold. The third phase involved the treatment of this
least susceptible life stage at the approximate mean
temperature of 1.1⬚C for various periods to determine
what length of treatment would be required. The
fourth phase was conducted to verify whether the
projected period of cold treatment at 1.1⬚C would
provide the required level of mortality.
Test Insects. B. invadens was reared at the ICIPE
mass rearing facility following the procedure de-
scribed by Ekesi et al. (2007). Adults that emerged
from puparia were fed on a 3:1 ratio of sugar and
enzymatic yeast hydrolysate ultrapure (USB Corpo-
ration, Cleveland, OH), and water on pumice gran-
ules. The viability of the eggs that the adult ßies pro-
duced for use in the various experiments were
monitored for each citrus infestation to ensure that the
results of the tests were not compromised by infertility
or poor egg viability. The viability was tested by col-
lecting Þve samples of eggs laid over a 1-h period
(Ekesi et al. 2007), and samples of 100 eggs were
placed on moist black cloth in a covered petri dish and
held at 28 ⫾ 2⬚C for 5 d to determine percentage of egg
hatch. The percentage of egg hatch on artiÞcial me-
dium would typically be close to 80% (Ekesi et al.
2007), but if it dropped to ⬍60% in any of the Þve
replicate petri dishes, the trials that used that batch of
eggs were discarded.
Test Fruit. Valencia oranges that were used for all
the experiments were of export quality purchased
from a local supplier in Nairobi. These oranges were
imported into Kenya from Egypt or South Africa and
met the required standards for maximum residue lev-
1181
els of plant protection products required by the Eu-
ropean Union for fresh citrus (Borrell-Fontelles and
Schmidt 2005). All the oranges were examined for the
presence of fruit ßies before they were used in tests,
and no damaged fruit were used.
Phase 1: Larval Development of B. invadens in
Oranges. To determine how long the different life
stages took to develop at 28⬚C, three replicates of the
following procedure were conducted using Valencia
oranges (69 mm in diameter). It was determined that
1 ml of B. invadens eggs was equivalent to 16,000 eggs,
and ⬇35 eggs were required per fruit in 0.025 ml of
liquid, so the eggs were diluted with the appropriate
amount of distilled water to provide this ratio. Counts
of the number of eggs in 10 aliquot samples per rep-
licate later showed that a mean number of 32 eggs per
fruit were used in the Þrst replicate and 33 in the
second and third replicates. The eggs (not older than
24 h) were inoculated into each orange on 14 August
2008 after Þrst removing the calyx and punching a hole
using a 6-mm-diameter cork borer into the stem end
of the fruit. Some Torula yeast (between 0.2 and 0.5 ml
of a mixture with water in a 1:2 ratio) also was injected
into the inoculation hole to provide additional nour-
ishment for the larvae. The inoculation procedure was
conducted as three batches of 130 fruit. Each batch
was treated as a replicate. After inoculation the hole
was plugged with sterilized cotton wool and later
sealed with hot wax. Each orange was placed in a
brown paper bag in a perforated plastic crate (Polyair
240 with dimensions 585 by 400 by 240 mm, Kenpoly
Manufacturers Ltd., Nairobi, Kenya). Approximately
75 fruit were placed in each crate with one bagged
fruit being placed in each crate before a second fruit
was placed in a crate. The crates of fruit were stored
in an incubation room (4,360 by 3,360 by 2,400 mm) at
28⬚C and positioned according to the replicate. The
temperature was maintained in the room by an air-
conditioning system and monitored with the use of
three different types of thermometers: a mercury
bulb, an alcohol bulb (both 305 mm, type BS1704-S,
Brannan [United Kingdom] laboratory thermometers
meeting ISO 1770:1981 requirements from Lowveld
Chemicals CC, Nelspruit, South Africa), and an elec-
tronic thermometer (SIKA, Kaufungen, Germany,
with accuracy ⱕ1% of range). From the Þrst day
through to the 15th day after inoculation, seven fruit
were dissected daily from each crate, and the number
of fruit ßy larvae in each instar determined. Dissected
larvae were carefully measured under a stereomicro-
scope with the assistance of Leica Application Suite
software, and at least six larvae per day were mounted
on slides and examined closely to determine what
instars were represented. Although the primary ob-
jective for conducting this study was to determine
how long it took for each instar to develop, a further
25 infested fruit per replicate were placed on sand and
the number of pupae determined daily by sieving the
sand from 19 August onward. The pupae were allowed
to develop and the number of adults emerging deter-
mined to gain more information about the life cycle of
B. invadens in citrus at 28⬚C.
1182 JOURNAL OF ECONOMIC ENTOMOLOGY
Phase 2: Determination of the Most Cold-Tolerant
Life Stage. The cold room (3,450 by 2,380 by 2,620
mm) used at ICIPE had a new Zanotti Uniblock SP-O
refrigeration unit Þtted that was well suited to the
required temperature range. The temperature probes
used were type “T” thermocouples with a tolerance of
0.5⬚C (Temperature Controls Pty., Ltd., Randburg,
South Africa). These were calibrated before each
treatment using the freezing point method where the
probes were immersed in melting ice, and the tem-
perature recorded when they reached equilibrium. A
thermometer immersed in the melting ice was used to
conÞrm the temperature. At least three calibration
runs were conducted, and the mean result for each
probe was used for correction purposes. Readings
were recorded hourly on a data logger (Grant SQ 2020
1F8, Monitoring and Control Laboratories Pty. Ltd.,
Johannesburg, South Africa). Thirty-three cartons
(400 by 300 by 270 mm) of export quality Valencia
oranges (69 mm in diameter) were acquired from
Dole South Africa Pty. Ltd., Bellville, South Africa, via
the Kenyan importer East Africa Growers, Nairobi for
both replicates. The fruit had been harvested in South
Africa ⬇3 wk earlier and kept in cold storage while
being transported and after arriving in Kenya. The Þrst
replicate started with inoculation on 13 October 2008,
and the second replicate started on 8 December 2008.
In each replicate, eggs (⬍24 h old) were inoculated
into 1,400 fruit as described above (35 per fruit), and
the fruit stored in paper bags that were placed in
perforated plastic crates as described for phase 1. One
group of 300 fruit was moved to the cold room (set for
a mean fruit pulp temperature of 1.1⬚C) on the day of
inoculation (egg stage), and the rest of the fruit were
held at 28⬚C. Fourteen temperature probes were in-
serted (30 mm in depth) into the pulp of oranges
scattered among the other fruit in the cold room, and
two other probes were used to monitor the incoming
and outgoing air temperatures in front of the circu-
lation fans. Whenever fruit was moved from the in-
cubation room into the cold room, it was moved 12 h
earlier than the time required for the start of the cold
treatment for the fruit to start cooling down from 28
to 1.1⬚C. Thus for the treatment that started 2 d after
inoculation, 300 fruit (containing Þrst instars) were
moved into the cold room 36 h after inoculation. A
third batch of 300 fruit (containing second instars)
was moved into the cold room 84 h after inoculation
for the 4-d treatment. After 6 d, the last batch of 300
fruit (third instars) was moved into the cold room. The
remaining 200 fruit (four batches of 50 fruit per life
stage group) were used as controls and held at 28⬚C
and then dissected 6 d after inoculation. Fifty fruit
from the egg stage group were removed from the cold
room on each of days 3, 5, 7, 9, 11, and 13 after
commencement of the cold treatment and held at 28⬚C
for 6 d. The fruit was then dissected and the number
of living and dead (no movement after prodding and
often with brown discoloration) larvae noted. Mor-
talities were corrected according to Abbott (1925).
Batches of 50 fruit from the Þrst-, second-, and third-
instar stage were removed at the same time intervals
Vol. 104, no. 4
as the egg stage group and held at 28⬚C for 4, 2, and 1 d,
respectively, before being dissected. Two replicates
were conducted and the susceptibilities of the differ-
ent life stages to the cold treatment were analyzed
with Probit analysis (Finney 1971) using the Proban
program (van Ark 1995) with 95% Þducial limits
(FLs).
Phase 3: Determining the Required Cold Period.
The Þrst replicate of phase 3 was started on 17 April
2009. B. invadens eggs were supplied from ICIPEÕs
culture and using export-quality Valencia oranges im-
ported from Dole South Africa Pty. Ltd., Bellville,
South Africa, by East Africa Growers, Nairobi, Kenya.
Eggs (54 per fruit) were inoculated into 4,320 Valencia
oranges with some Torula yeast and the fruit held at
28⬚C until the fruit ßy reached third instars (6 d). A
cork borer (6 mm in diameter) was used to make a
hole for inoculation in the depression left after re-
moving the calyx. The holes were later plugged with
cotton wool and sealed with wax. Three hundred and
sixty control oranges were dissected and inspected
through 5-diopter head loupes or stereomicroscopes
on 23 April to determine the number and survival of
third instars. The remaining 3,960 oranges were trans-
ferred into the cold room on the same day. The 55
crates of fruit were arranged in 11 stacks, Þve crates
high, and a gap of 5 cm was kept around all crates to
assist air ßow. There was a gap of at least 50 cm
between any crates and the walls of the cold room, and
the room could contain at least 100 crates. Tempera-
tures were recorded hourly as before using type “T”
thermocouples (described above); 14 in the pulp of
oranges (30 mm in depth) and two to monitor the
incoming and outgoing air temperatures in front of the
circulation fans.
The room temperature was brought down as
quickly as possible from ambient (⬇26⬚C), but due to
the large amount of warm fruit it took almost 3 d for
half of the 14 probes in the fruit to reach the required
temperature or below (a similar trend is shown in Fig.
3), which was when the cold treatment of 1.1⬚C was
deemed to have started. Batches of 660 oranges were
then removed from the cold room after 3, 5, 7, 9, 11,
and 13 d of treatment and kept at 28⬚C for 24 h before
being cut. Oranges were inspected under magniÞca-
tion to determine the number of surviving larvae.
The second replicate of phase 3 was initiated on 2
July 2009 with inoculation of 4,300 export-quality Va-
lencia oranges from the same source as before using 58
eggs per fruit as described above. After having been
incubated at 28⬚C for 6 d, 378 fruit were cut to deter-
mine infestation and mortality levels, and the remain-
ing fruit were moved into the cold room on 8 July. The
fruit reached the required temperature on 11 July, and
thereafter 660 fruit were removed for cutting and
determination of mortality at the required intervals as
described above.
The data from phase 3 were subjected to Probit
analysis using the Proban program (van Ark 1995) to
determine the length of cold treatment required.
August 2011 GROUT ET AL.: B. invadens COLD SUSCEPTIBILITY IN ORANGES 1183

Phase 4: Verification of Temperature–Period Com- Table 1. Phase 1: Larval development (based on measure-
bination. The Þrst replicate of phase 4 to evaluate a ment) at 28°C with time after inoculation for three replicates of 105
fruit each inoculated with 32 (replicate 1) or 33 (replicates 2 and
16-d period at 1.1⬚C was initiated on 20 August 2009 3) eggs
when 2,500 export-quality Valencia oranges imported
from Dole South Africa Pty. Ltd. by East Africa Grow- SE of Instar
Days after Mean size Total no.
ers were inoculated with ⬇50 eggs each according to inoculation (mm)
replicate
larvae
ratio
the technique described above. The fruit were incu- means (%)a
bated at 28⬚C until 26 August when 500 control fruit 1 0 No larvae
were cut to determine larval infestation levels and 2 2.90 0.20 59 100:0:0:0
3 4.81 0.14 71 93:7:0:0
natural mortality, and the remaining fruit was moved 4 7.58 0.75 113 0:87:13:0
into the cold room. The 40 crates were arranged in 5 9.01 0.30 129 0:30:70:0
eight stacks, Þve crates high, with 5-cm gaps around 6 9.45 0.09 134 0:9:91:0
the blocks. Fourteen temperature probes were again 7 9.72 0.10 149 0:0:99:1
8 9.96 0.09 146 0:0:97:3
inserted into oranges scattered among the other fruit 9 10.03 0.15 145 0:0:97:3
in the cold room, and two other probes were used to 10 10.19 0.19 93 0:0:78:22
monitor the incoming and outgoing air temperatures. 11 9.85 0.13 95 0:0:88:12
Readings were recorded hourly using the same data 12 10.19 0.07 124 0:0:85:15
13 10.01 0.03 116 0:0:87:13
logger as described above. The fruit reached the re- 14 9.80 0.06 103 0:0:80:20
quired temperature after 2 d on 28 August. On 13 15 10.28 0.08 103 0:0:80:20
September, the fruit was transferred to the 28⬚C room Total 1580
after 15 d and 23 h. The disinfestation period was a
intentionally terminated 1 h early to prevent any vari- First:second:third:puparia.
ance in interpretation of temperature records from
extending the period beyond 16 d. The next day, the imported from Dole South Africa Pty. Ltd. by East
fruit was cut once it had reached ambient tempera- Africa Growers, each receiving 55 eggs. The fruit was
ture. Numbers of larvae alive or dead were recorded incubated at 28⬚C until 21 April when 544 control fruit
in each fruit. were cut, and the remaining fruit were transferred to
The second replicate of phase 4 was inoculated on the cold room. The crates were arranged as before.
12 September by using another 2,500 export-quality The target pulp temperature of 0.5⬚C (SEM ⫾ 0.0) was
Valencia oranges from Swaziland purchased from East reached on 24 April after 69 h cooling, and fruit were
Africa Growers and 50 eggs per fruit. The fruit was transferred to the incubation room on 10 May after
incubated at 28⬚C until 18 September when 500 con- 15 d and 23 h. After reaching ambient temperature, the
trol fruit were cut, and the remaining fruit were trans- fruit were cut, and numbers of live and dead larvae
ferred to the cold room and the crates arranged as were determined.
before. The required fruit pulp temperature of 1.1⬚C The third replicate at a mean pulp temperature of
was reached on 21 September, and the fruit was re- 0.5⬚C (probe tolerance ⫾0.5) was inoculated on 14
moved on 7 October after 15 d and 23 h. After being May 2010 with 3,552 export-quality Valencia oranges,
held at 28⬚C for 24 h, the fruit was cut, and numbers imported from Dole South Africa Pty. Ltd. by East
of live and dead larvae were determined. Africa Growers, each receiving 44 eggs. The fruit were
Surviving larvae were found in the second replicate incubated at 28⬚C until 20 May when 520 control fruit
at 1.1⬚C (SEM ⫾ 0.0 and probe tolerance ⫾0.5). Ex- were cut, and the remaining fruit were transferred to
tending the treatment period beyond 16 d would be the cold room. The crates were arranged as described
impractical for the shipping of fruit to Japan, so rather above. The target temperature of 0.5⬚C (SEM ⫾ 0.0)
than conducting the third planned replicate using the was reached on 22 May after 52 h, and fruit were
same conditions a new phase 4 series at an approxi- transferred to the incubation room on 7 June after 15 d
mate mean pulp temperature of 0.5⬚C for 16 d was and 23 h. After reaching ambient temperature, the
initiated. Inoculation of the Þrst replicate of this series fruit were cut, and numbers of live and dead larvae
took place on 4 October 2009, by using 2,830 imported were determined.
Valencia oranges and 44 eggs per fruit. The fruit was While fruit were being cut in this second phase 4
incubated at 28⬚C until 10 October (6 d) when 585 experiment, they were cursorily examined for the
control fruit were cut and the remaining fruit were presence of any signs of chilling injury.
transferred to the cold room. The crates were ar-
ranged as described above. The new target pulp tem-
Results
perature of 0.5⬚C (SEM ⫾ 0.0 and probe tolerance
⫾0.5) was reached on 13 October after 65 h, and fruit Phase 1: Larval Development of B. invadens in
was transferred to the incubation room after 15 d and Oranges. The results of the larval development stud-
23 h on 29 October. After reaching ambient temper- ies at 28⬚C (Table 1) showed that Þrst instars appeared
ature, the fruit was cut, and numbers of live and dead 2 d after inoculation and that most second instars
larvae were determined. appeared 4 d after inoculation. Most larvae had
The second replicate at a mean pulp temperature of reached the third instar by 6 d after inoculation. Com-
0.5⬚C (probe tolerance ⫾0.5) was inoculated on 15 pared with the short larval life stages, the formation of
April 2010 with 2,650 export-quality Valencia oranges, puparia took place over a long time (21 d), perhaps as
1184 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 4

35 outgoing air temperature dropped to 1.0⬚C within 2 h.

Puparia
The natural mortality of B. invadens in oranges be-
30
Adults tween hatching and the third instar was high (Table
25
2), indicating that orange is not an ideal host. Seven
days after inoculation, there was 100% mortality of
Numbers per day

20 eggs in both replicates (Table 2), and the same results

15
10
5
0 licate, whereas for third instars the corrected mortal-
5 10 15 20 25 30
Days after inoculation
Fig. 1. Phase 1: total numbers of B. invadens puparia
formed and adults eclosing per day after larvae emerged from
oranges on sand at 28⬚C.
a survival strategy or due to the larvae having difÞculty
in exiting the fruit (Fig. 1). The Þrst larvae left the fruit
to pupate 7 d after inoculation. The Þrst adults eclosed
17 d after inoculation and most adults eclosed between
20 and 26 d after inoculation (Fig. 1).
Phase 2: Determination of the Most Cold-Tolerant
Life Stage. When the Þrst few crates of 300 fruit were
moved into the cold room the pulp temperature was
lowered from 24⬚C to 1.1⬚C within 19 h while the
were obtained when larvae began treatment in the
Þrst instar. With only two results from each of these
life stages having mortalities ⬍100%, Probit analysis
was not possible. The corrected mortalities of second
instars after having been exposed for 7 d were 99.73%
for the Þrst replicate and 99.29% for the second rep-
35 40 ities were 97.83 and 98.68%, respectively. With Probit
analysis, these differences between second and third
instars were signiÞcant (P ⬍ 0.05) in the Þrst replicate
(Table 3) but not in the second replicate. However,
the prediction for the lethal time (LT) 99.9% was
higher for third instars than for second instars in both
replicates. These results therefore suggested that the
third instar was more tolerant to the cold treatment,
so this instar was used in further evaluations.
Phase 3: Determining the Required Cold Period.
Although the results of the Þrst replicate seemed to
show a steady increase in mortality with time (Table
4; Fig. 2) and had an R2 value of 0.9804, the Probit
analysis showed extreme variation with a high G value
of 0.28 and wide 95% FLs. The prediction of 19.3 d for
99.9% mortality (Table 4) was therefore considered to

Table 2. Phase 2: Mortality resulting from exposure of different life stages to 1.1°C in oranges for different periods of time in addition
to 6 d at 28°C (egg mortality in media is typically 20%)

Replicate 1 Replicate 2
Exposure
Stage No. eggs No. live Mortality No. eggs No. live Mortality
period (d)
treated larvae recovered (%) treated larvae recovered (%)
Eggs 0 Control 1,785 333 81.34 1,961 271 86.18
3 1,750 126 92.80 1,850 80 95.68
5 1,750 16 99.09 1,924 11 99.43
7 1,750 0 100.00 1,924 0 100.00
9 1,750 0 100.00 1,924 0 100.00
11 1,750 0 100.00 1,850 0 100.00
13 1,820 0 100.00 1,924 0 100.00

First instar 0 Control 1,750 347 80.17 1,961 284 85.52

3 1,750 96 94.51 1,850 71 96.16
5 1,785 14 99.22 1,924 15 99.22
7 1,750 0 100.00 1,924 0 100.00
9 1,785 0 100.00 1,924 0 100.00
11 1,785 0 100.00 1,924 0 100.00
13 1,750 0 100.00 1,924 0 100.00

Second instar 0 Control 1,750 364 79.20 1,961 291 85.16

3 1,750 143 91.83 1,887 114 93.96
5 1,750 35 98.00 1,924 15 99.22
7 1,785 1 99.94 1,887 2 99.89
9 1,750 0 100.00 1,924 2 99.90
11 1,750 0 100.00 1,924 0 100.00
13 1,785 0 100.00 1,924 0 100.00

Third instar 0 Control 1,785 361 79.78 1,887 380 79.86

3 1,785 235 86.83 1,924 131 93.19
5 1,750 78 95.54 1,924 29 98.49
7 1,820 8 99.56 1,887 5 99.74
9 1,750 2 99.89 1,924 0 100.00
11 1,785 0 100.00 1,887 0 100.00
13 1,750 0 100.00 1,998 0 100.00
August 2011 GROUT ET AL.: B. invadens COLD SUSCEPTIBILITY IN ORANGES 1185

Table 3. Phase 2: Expected treatment periods required for 99.9% mortality of second and third instars at 1.1°C

Instar Replicate Days SE Upper FL Lower FL

Second Firsta 9.83 1.834 Not possible Not possible
Third Firsta 11.19 0.641 12.71 10.11
Second Secondb 8.74 0.718 10.62 7.62
Third Secondb 10.92 1.023 13.61 9.32

a
Comparison of second and third instar lines: elevations were signiÞcantly different (F ⫽ 15.951; df ⫽ 1, 4; P ⫽ 0.0160).
b
Comparison of second and third instar lines: elevations were not signiÞcantly different (F ⫽ 0.008; df ⫽ 1, 3; P ⫽ 0.9340).

be unreliable due to the variation (although the later Discussion

phase 4 results with greater replication indicated that
this result was quite realistic). With the second rep-
licate, results (Table 4) were less variable and the
Probit prediction for 99.9% mortality was 15.1 d. The
overall trend therefore seemed similar to previous
results with Mediterranean fruit ßy in Valencia or-
anges (Ware et al. 2006, Grout et al. 2011) for which
16 d at a mean temperature of 1⬚C was found to be
effective. We therefore decided to verify whether a
16-d treatment at a mean fruit pulp temperature of
1.1⬚C would provide mortality above the 99.99% level
with 95% conÞdence levels.
Phase 4: Verification of the Temperature–Period
Combination. The Þrst replicate showed no survivors
of 16,617 third instars, but in the second replicate three
survivors were found in 23,536 larvae (Table 5). Two
survivors were from one fruit and one survivor from
another fruit. The two fruit were cut by different
people and the fruit came from different crates in
different parts of the cold room, so the survival could
not be attributed to a hot spot in the room. The
temperature probes showed no unusually high tem-
peratures (Table 5) that could have explained the
survival, so we had to conclude that this cold treat-
ment was inadequate.
In the three replicates of the phase 4 trials at 0.5⬚C
(mean of 0.5⬚C, with the mean of hourly maximum
temperatures of 0.9⬚C) (Table 5; Fig. 3), no survivors
were found in a total of 65,752 treated larvae. The
numbers of larvae treated in each of the three repli-
cates were 16,437, 17,847, and 31,468. This treatment
therefore exceeded the required minimum assurance
of 99.99% mortality at the 95% conÞdence level and
also passed the Probit-9 mortality level. No signs of
chilling injury were observed.
Although B. invadens currently occurs in tropical
areas its minimum developmental threshold for larvae
is 9.4⬚C (Rwomushana et al. 2008b), which is lower
than that of South African C. capitata larvae found to
be 10.5⬚C by Grout and Stoltz (2007). Our work can-
not provide conclusive evidence that B. invadens is
more cold tolerant than C. capitata in citrus, because
in these trials B. invadens was exposed to a slightly
higher temperature than used by Grout et al. (2011)
for C. capitata when using a larval end point. However,
earlier research with a pupariation end point showed
that a mean fruit temperature of 1.0⬚C for 16 d caused
100% mortality of C. capitata in oranges (Hill et al.
1988) and 1.0⬚C for 14 d was adequate for complete
control of C. capitata in lemons, Citrus limon (L.)
Burm.f., when a pupariation end point was used (Jes-
sup et al. 1993). The energy required to form a pupa
would probably result in some mortality of weak sur-
viving larvae but without a direct comparison we can-
not speculate about the equivalence of this mortality
in terms of days of treatment or lower treatment tem-
perature. Recent results from small-scale (no survi-
vors from 104 larvae treated for 11 d) incubator studies
in media showed that B. invadens was no more cold
tolerant than C. capitata when treated at a mean tem-
perature of 0.94⬚C (Hallman et al. 2011). Our exposure
of 21,784 third instars to 1.1⬚C for 11 d resulted in 27
survivors (Table 4), so perhaps if the comparison of
Hallman et al. (2011) was repeated with larger num-
bers a different result may be obtained.
Research by Rwomushana et al. (2008a) showed
that oranges were not as good a host as mangoes,
bananas (Musa spp.), or papayas, Carica papaya L., for
B. invadens and that both fecundity and fertility were

Table 4. Phase 3: mortalities of B. invadens third instars in Valencia oranges when treated for different periods at 1.1°C

Replicate 1a Replicate 2b
Days
Total larvae Live larvae Mortality (%) Total larvae Live larvae Mortality (%)
0 Check 4,924 4,219 14.3 5,101 4,151 18.6
3 9,190 3,977 56.7 9,358 4,342 53.6
5 10,300 2,791 72.9 10,247 2,089 79.6
7 11,045 1,141 89.7 9,646 598 93.8
9 11,024 212 98.1 9,910 84 99.2
11 11,110 9 99.92 10,674 18 99.8
13 11,196 1 99.99 11,253 0 100.0

a
Probit analysis for an expected mortality of 99.9% predicted a period of 19.3 d with SE of 4.337, upper and lower Þducial limits of 84.71
and 11.89, and G value of 0.28.
b
Probit analysis for an expected mortality of 99.9% predicted a period of 15.1 d with SE of 1.546, upper and lower Þducial limits of 23.47
and 11.60, and G value of 0.076.
1186
100
90
Corrected mortality (%)
80
70
60
50
40
0.4 0.6
Log 10 days
Fig. 2. Mortality of third instars in oranges for both rep-
licates of phase 3 after applying AbbottÕs correction for the
period in days expressed as log10.
signiÞcantly lower in oranges than in these other hosts.
This explains why the mortality between egg hatch
and third instar in the controls was usually close to 80%
(Table 2), even though steps were taken to ensure that
only eggs of good viability were used for inoculation.
This result is not very different to C. capitata in or-
anges where mortality between egg hatch and second
instar has been shown to be 72% (Grout et al. 2011).
Other studies (Hill et al. 1988, Jessup et al. 1993) did
not determine how many eggs were oviposited in fruit
by C. capitata, so they could not estimate mortality
based on egg numbers, but Sproul (1976) recovered
⬇20 C. capitata pupae per ÔGranny SmithÕ apple after
injecting 150 Ð200 eggs per apple, which results in a
mortality of at least 87%.
The time taken to get the fruit pulp temperature
down from ⬇26⬚C to 0.5⬚C took longer than 48 h in two
of the three replicates in phase 4 (Fig. 3), but this is
not considered unusual in comparison with commer-
cial situations where large consignments of fruit in
cartons need to be cooled to similar temperatures. If
a cold treatment schedule was based on a much
shorter precooling period, its relevance to what would
happen in practice could be questioned.
The above-mentioned results and experimental
treatment conditions provide the basis for establishing
a treatment protocol for practical use in phytosanitary
regulation of trade with B. invadens host material such
Table 5. Phase 4: efficacy of 1.1°C for 16 d in two replicates and the efficacy of 0.5°C for 16 d in three replicates against B. invadens
third instars, with mean fruit pulp temperatures
Statistic
No. fruit
No. live larvae
Total larvae
Larval mortality (%)
Mean of hourly avg. temp, ⬚C (SEM)
Mean of hourly max. temp, ⬚C (SEM)
Mean of hourly min. temp, ⬚C (SEM)
a
Probit-9 requirement is 99.9968% mortality.
JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 4
as citrus. The critical point in establishing such a treat-
R2 = 0.9447 ment protocol is selection of the upper threshold tem-
R2 = 0.9804
perature. Treatment protocols set a minimum expo-
sure period and a maximum temperature (upper
threshold). For adequate assurance of efÞcacy during
trade with potentially infested host material it is nec-
essary to ensure that the product has been exposed to
temperatures at or below the upper threshold for at
least the minimum exposure period. Whereas temper-
ature ßuctuations are minimized during the validating
Replicate 1 experimental work, under conditions of practical ap-
Replicate 2
plication to a large bulk of host material (such as
export shipment of fruit), temperature ßuctuations
0.8 1.0
are far greater than under experimental conditions
(Tanner and Amos 2003). Consequently, practical
product treatment occurs at temperatures consider-
ably below the upper threshold to avoid exceeding this
threshold and invalidating the treatmentÕs application.
Under practical shipping conditions the maximum
fruit pulp temperature sensors are not permitted to
exceed a temperature of at least 0.5⬚C below the upper
threshold, and delivery air temperature is generally set
at a further 0.5 to 1⬚C below this target temperature
(i.e., 1 to 1.5⬚C below the upper threshold; PPECB
2009). Therefore, from a phytosanitary security per-
spective, it is generally safe to set the maximum pro-
tocol temperature above the mean of the treatment
condition maintained in the supporting experiments.
Using the mean of the average hourly maximum probe
readings during the experiment provides a conserva-
tive measure for the upper threshold to be deÞned in
a practical treatment protocol.
Direct comparison between these results and the
USDAÐAPHIS treatment protocol of T107-a for C.
capitata and C. rosa is difÞcult, because trial proce-
dures deployed will not have been the same and the
calculations used to convert previous trial conditions
into the T107-a treatment protocol are unknown.
Nonetheless, considering that there were a few sur-
viving larvae in replicate two of phase 4 at a treatment
midpoint of 1.1⬚C for 16 d, it seems that the T107-a
protocol (1.11⬚C for 14 d or 1.67⬚C for 16 dÑ being the
upper threshold temperatures, as opposed to midpoint
treatment temperatures) might not always provide
adequate assurance of efÞcacy (when measuring lar-
val survival), depending on how closely the practical
application of the treatment is allowed to match the
Totals from two replicates Totals from three replicates
Control at 28⬚C 16 d at 1.1⬚C Control at 28⬚C 16 d at 0.5⬚C
1,025 4,108 1,649 7,379
9,619 3 11,059 0
10,075 40,153 12,112 65,752
4.5 99.9925a 8.1 100.0
1.1 (0.0) 0.5 (0.0)
1.4 (0.0) 0.9 (0.0)
0.8 (0.1) 0.2 (0.0)
August 2011 GROUT ET AL.: B. invadens COLD SUSCEPTIBILITY IN ORANGES 1187

these results support the establishment of the follow-

25 Mean hourly
Max hourly
ing phytosanitary treatment protocol for B. invadens
Min hourly disinfestation in citrus fruit: fruit pulp to be main-
20
tained at temperatures of 0.9⬚C or lower for 16 con-
a secutive days.
15

10

Acknowledgments
5
We thank Mike Hennessey for advice and anonymous
0
reviewers for helpful comments, numerous casual laborers
0 48 96 144 192 240 288 336 384 432 for evaluating the many treatments and those responsible for
maintaining the B. invadens culture at ICIPE. This research
25 Mean hourly
was funded in part by the Citrus Growers Association of
Max hourly Southern Africa and support from the ICIPE core funds.
Min hourly
20
Temperature (C)

b
15 References Cited
10
Abbott, W. S. 1925. A method of computing the effective-
ness of an insecticide. J. Econ. Entomol. 18: 265Ð267.
5
[APHIS] U.S. Animal and Plant Health Inspection Service.
2010. U.S. Animal and Plant Health Inspection Service,
0
Treatment manual 07/2010-47 Chapter 5-2. T100-Sched-
0 48 96 144 192 240 288 336 384 432 ules for Fruit, Nuts, and Vegetables. (http://www.aphis.
usda.gov/import_export/plants/manuals/ports/treatment.
Mean hourly
shtml).
25
Max hourly Borrell-Fontelles, J. P., and N. Schmidt. 2005. Regulation
Min hourly (EC) No 396/2005 of the European Parliament and of the
20
Council of 23 February 2005 on maximum residue levels
c of pesticides in or on food and feed of plant and animal
15
origin and amending Council Directive 91/414/EEC. Of-
Þcial J. European Union L70: 1Ð16.
10
Drew, R., K. Tsuruta, and I. White. 2005. A new species of
pest fruit ßy (Diptera: Tephritidae: Dacinae) from Sri
5
Lanka and Africa. Afr. Entomol. 13: 149 Ð154.
Ekesi, S., and E. Muchugu. 2007. Tephritid fruit ßies in Af-
0
0 48 96 144 192 240 288 336 384 432
rica Ð fact sheets of some economically important species,
B-10. In S. Ekesi and M. K. Billah (eds.), A Þeld guide to
Hours
the management of economically important tephritid
Fig. 3. Hourly maximum (dotted line), hourly minimum fruit ßies in Africa, 2nd ed. EICIPE Science Press, Nairobi,
(dashed line), and hourly mean (solid line) temperatures of Kenya.
fruit pulp in three replicates during cold treatment. Long- Ekesi, S., M. K. Billah, P. W. Nderitu, S. A. Lux, and I.
term mean temperatures were as follows. (a) Replicate 1: Rwomushana. 2009. Evidence for competitive displace-
mean of hourly averages, 0.5⬚C; mean of hourly minima, ment of the mango fruit ßy, Ceratitis cosyra by the inva-
0.1⬚C; mean of hourly maxima, 0.9⬚C. (b) Replicate 2: mean sive fruit ßy, Bactrocera invadens (Diptera: Tephritidae)
of hourly averages, 0.5⬚C; mean of hourly minima, 0.2⬚C; on mango and mechanisms contributing to the displace-
mean of hourly maxima, 0.9⬚C. (c) Replicate 3: mean of ment. J. Econ. Entomol. 102: 981Ð991.
hourly averages, 0.5⬚C; mean of hourly minima, 0.3⬚C; mean Ekesi, S., P. W. Nderitu, and C. L. Chang. 2007. Adaptation
of hourly maxima, 0.8⬚C. to and small-scale rearing of invasive fruit ßy Bactrocera
invadens (Diptera: Tephritidae) on artiÞcial diet. Ann.
Entomol. Soc. Am. 100: 562Ð567.
Ekesi, S., P. W. Nderitu, and I. Rwomushana. 2006. Field
upper threshold temperatures stipulated in the pro- infestation, life history and demographic parameters of
tocol. When considering that the objective of such a the fruit ßy Bactrocera invadens (Diptera: Tephritidae) in
treatment is to prevent the treated commodity from Africa. Bull. Entomol. Res. 96: 379 Ð386.
providing a pathway for the establishment of the or- [FAO] Food and Agriculture Organisation of the United
ganism in new areas, evaluating larval survival is a Nations. 2006. Citrus fruit: fresh and processedÑannual
conservative estimate of the treatmentÕs efÞcacy in statistics 2006. Food and Agriculture Organisation of the
preventing emergence of fertile adults. Therefore, United Nations. (http://www.fao.org/es/esc/common/
T107-a may still provide adequate phytosanitary se- ecg/243/en/bull2006.pdf).
curity but could complicate regulatory application if Finney, D. J. 1971. Probit analysis. Cambridge University
Press, Cambridge, United Kingdom.
low levels of larval survival were to be detected during Follett, P. A., and L. G. Neven. 2006. Current trends in
posttreatment inspection. quarantine entomology. Annu. Rev. Entomol. 51: 359 Ð
Taking the above-mentioned information into con- 385.
sideration, and that the mean of the average hourly Grout, T. G., P. R. Stephen, J. H. Daneel, and V. Hattingh.
maximum probe readings was 0.9⬚C, we conclude that 2011. Cold disinfestation of Ceratitis capitata (Diptera:
1188 JOURNAL OF ECONOMIC ENTOMOLOGY
Tephritidae) in oranges using a larval endpoint. J. Econ.
Entomol. 104: 1174 Ð1179.
Grout, T. G., and K. Stoltz. 2007. Developmental rates at
constant temperatures of three economically important
Ceratitis spp. (Diptera: Tephritidae) from southern Af-
rica. Environ. Entomol. 36: 1310 Ð1317.
Hallman, G. J., S. W. Myers, A. J. Jessup, and A. Islam. 2011.
Comparison of in vitro heat and cold tolerances of the
new invasive species Bactrocera invadens (Diptera: Te-
phritidae) with three known tephritids. J. Econ. Entomol.
104: 21Ð25.
Hill, A. R., C. J. Rigney, and A. N. Sproul. 1988. Cold storage
of oranges as a disinfestation treatment against the fruit
ßies Dacus tryoni (Froggatt) and Ceratitis capitata (Wie-
demann) (Diptera: Tephritidae). J. Econ. Entomol. 81:
257Ð260.
[IPPC]. 2010. International Plant Protection Convention:
South Africa. (https://www.ippc.int/index.php?id⫽
nppoZAF).
Jessup, A. J., C.P.F. De Lima, C. W. Hood, R. F. Sloggett, A. M.
Harris, and M. Beckingham. 1993. Quarantine disinfes-
tation of lemons against Bactrocera tryoni and Ceratitis
capitata (Diptera: Tephritidae) using cold storage. J.
Econ. Entomol. 86: 798 Ð 802.
Lux, S. A., R. S. Copeland, I. M. White, A. Manrakhan, and
M. K. Billah. 2003. A new invasive fruit ßy species from
the Bactrocera dorsalis (Hendel) group detected in East
Africa. Insect Sci. Appl. 23: 355Ð361.
Mwatawala, M., M. De Meyer, R. H. Makundi, and A. P.
Maerere. 2006. Seasonality and host utilization of the
invasive fruit ßy, Bactrocera invadens (Dipt., Tephritidae)
in central Tanzania. J. Appl. Entomol. 130: 530 Ð537.
Ndiaye, M., E. O. Dieng, and G. Delhove. 2008. Population
dynamics and on-farm fruit ßy integrated pest manage-
Vol. 104, no. 4
ment in mango orchards in the natural area of Niayes in
Senegal. Pest Manag. Hortic. Ecosyst. 14: 1Ð 8.
[PPECB] Perishable Products Export Control Board. 2009.
Procedures for in-transit cold treatment citrus fruit ship-
ments to Japan HP39 Rev. 01. Perishable Products Export
Control Board, Cape Town, South Africa.
Rwomushana, I., S. Ekesi, I. Gordon, and C.K.P.O. Ogol.
2008a. Host plants and host plant preference studies for
Bactrocera invadens (Diptera: Tephritidae) in Kenya, a
new invasive fruit ßy species in Africa. Ann. Entomol. Soc.
Am. 101: 331Ð340.
Rwomushana, I., S. Ekesi, C.K.P.O. Ogol, and I. Gordon.
2008b. Effect of temperature on development and sur-
vival of immature stages of Bactrocera invadens (Diptera:
Tephritidae). J. Appl. Entomol. 132: 832Ð 839.
Sproul, A. M. 1976. Disinfestation of Western Australian
Granny Smith apples by cold treatment against the egg
and larval stages of the Mediterranean fruit ßy, Ceratitis
capitata (Wiedemann). Austral. J. Exp. Agric. Anim.
Husb. 16: 280 Ð285.
Tanner, D. J., and N. D. Amos. 2003. Temperature variabil-
ity during shipment of fresh produce. Acta Hortic. 599:
193Ð203.
van Ark, H. 1995. Introduction to the analysis of quantal
response. Agricultural Research Council, Agrimetrics In-
stitute, Pretoria, South Africa.
Ware, A. B., B. Tate, J.-H. Daneel, P. Stephen, and R. Beck.
2006. Sensitivity of Mediterranean fruit ßy eggs and lar-
vae in lemons, grapefruit and oranges to cold treatment
of 1⬚C, pp. 100 Ð106. In: CRI Group Annual Research
Report for 2005. Citrus Research International, Nelspruit,
South Africa.
Received 2 December 2010; accepted 11 May 2011.