Content uploaded by Stephen Morse
Author content
All content in this area was uploaded by Stephen Morse on Dec 24, 2013
Content may be subject to copyright.
Ou tlo ok on AG RICULTUR E Vol 32, No 2, 2003, pp 123–128 123
Bt cotton, pesticides,
labour and health
A case study of smallholder
farmers in the Makhathini Flats,
Republic of South Africa
Richard Bennett, T. Joseph Buthelezi,
Yousouf Ismael and Stephen Morse
Abstract: This paper describes some of the results of a detailed farm-level survey of
32 small-scale cotton farmers in the Makhathini Flats region of South Africa. The
aim was to assess and measure some of the impacts (especially in terms of savings in
pesticide and labour as well as benefits to human health) attributable to the use of
insect-tolerant Bt cotton. The study reveals a direct cost benefit for Bt growers of
SAR416 ($51) per hectare per season due to a reduction in the number of insecticide
applications. Cost savings emerged in the form of lower requirements for pesticide,
but also important were reduced requirements for water and labour. The reduction
in the number of sprays was particularly beneficial to women who do some spraying
and children who collect water and assist in spraying. The increasing adoption rate
of Bt cotton appears to have a health benefit measured in terms of reported rates of
accidental insecticide poisoning. These appear to be declining as the uptake of Bt
cotton increases. However, the understanding of refugia and their management by
local farmers are deficient and need improving. Finally, Bt cotton growers emerge as
more resilient in absorbing price fluctuations.
Keywords: insecticide; health; smallholder farmers; Bt cotton; South Africa
Richard Bennett is with the Agricultural Economics and Policy Research Group, Department of
Agricultural and Food Economics, University of Reading, Reading, Berkshire RG6 6AR, UK.
Tel: +44 118 9316478. E-mail: r.m.bennett@reading.ac.uk. Yousouf Ismael is also with the
Department of Agricultural and Food Economics, University of Reading. Tel: +44 118 9318971.
E-mail: y.ismael@reading.ac.uk. T. Joseph Buthelezi is Chairman of Ubongwa District, Makhathini
Flats, Republic of South Africa. Stephen Morse is with the Department of Geography, University of
Reading. Tel: +44 118 9318736. E-mail: s.morse@reading.ac.uk.
The cotto n crop is suscep tible t o damage fro m a range of
insect pests such as the bo llworm complex, jassids and
aphids (M anuna, 1997). For this reason, la rge quantities of
the most toxic pesticides in the w orld are used in growing
cotton. Fo r example, 2 4% of the insecticides sold in the
world ma rket in 1994 were used on the 2.4% a rable land
area used for gro wing cotton (Myers, 2000). As pests,
particularly the bollworm complex, can severely reduce
yield and increase costs, then profitability declines and
farming risks increase, particularly for the small-scale and
reso urce-poor farmers. In additio n, the use o f some of
these pesticides can be particula rly da ngerous to farmers
and their worke rs (Londo n, 1998) as well a s to the
env ironment (Bet z
et al, 2000; Rother, 2000; Wilkins
et al,
2000; Yousefi, 2000). The lack o f training in the use of
pesticides, protective equipment and adequa te storage
124 O ut loo k o n A GRICU LTURE Vol 3 2, N o 2
Bt cotton, pesticides , labour and health
facilities are some of the contributory facto rs that make
pesticides partic ularly dangerous in small-sca le
production. Myers (2000) noted that in developing
countries pesticides we re the cause of as many as 14% of
work-related injuries , a nd that 1 0% of these inju ries led t o
fatality.
The introduction of insect-tolerant
Bacillus thuringiensis
(Bt) cotton should in princip le reduce the quantity of
pesticide used, and hence reduce production costs
(pesticide and labour) whilst improving farm workers ’
saf ety and the e nvironment. Savings in labour could free
up more time for farmers and their fa milies and at the
sam e t ime reduce farming risks. The
Bt
gene in cotton is
designed to protect against b ollworm complexes, ie the
three pests: tobacco budworm, cotton bo llwo rm an d p ink
bollworm (Monsanto, 2002). Several studies (Wier
et al,
1998; Edg e
et al, 2001), m ainly in the developed countries,
hav e c onfirmed t hat the use of
Bt co tto n reduces insecti-
cide c osts. Furthermore, numero us studies carried out in
the USA, Australia, Chin a, Mexico and Spain have
dem onstrated an overall reduction in the number of
sprays for lepido pteran pests (Stark, 1997; Mu llins a nd
Mills, 1999; Novillo
et al,
1999; Xia
et al, 1999; Benedict
and Altman, 2 001 ). Studies in developing c ountries by
Traxler
et al (2002) in Mexico, P ray
et al (2001) in China,
and Ismael
et al (2001, 2002a, 200 2b) in South Africa
confirmed the fact t hat the introduction of
Bt co tton does
indeed reduce insecticide costs and the number of
inse cticide applications per s easo n.
The aim of the research describe d in this paper w as to
test these assumed benefits in terms of pe stic ide and
lab our saving by measu ring a nd assessing the effect of
Bt co tton on pesticide use a mongst 32 smallholder
farmers in th e Makhathini reg ion of the Rep ublic o f
South Africa. Also inclu ded wa s an attemp t to gauge
possible benefits to human health. This study fo rms part
of a larger programme on
Bt cotton being co nducted by
the University o f Reading (Ismael
et al, 2001, 2002a,
2002b).
Agriculture is the most important source of income in
the Makha thini area , a nd rural households cultivate land
(typically one to three ha) allocated to them by their trib al
chiefs. The major crops in the are a are beans, maize and
cotton. The latt er usu ally occupies most of the farm and is
grown as a com merc ial cro p. P lant ing takes place fro m
mid-October t o mid-Decemb er and harvesting from mid-
May to mid-June. The main reason for growing cotton
(besides c ash) is that the crop need s less intensive
management than maize or be ans and c an survive
flu ctuating weather conditions. A depleted labour force
(du e to male migratio n to tow ns), low capitalization
(hence heavy reliance on credit), climatic constra ints (eg
erra tic rainfall patterns) and pest attack are the main
constraints to fa rming in the reg ion.
Bt co tton was released in M akhathini in the 1998
/
99
growing seaso n. A Mo nsanto report (Bennett, 2002) shows
tha t in 1998
/
99, there were only 75 adopters of the n ew
Bt
cotton (NUCOTN 37 B) in Makhathini, grow ing less tha n
200 hectares. In 1999
/
00, this rose to 411 ad opters with a
litt le under 70 0 hectares, a nd in 2000
/
01, to 1,184 adopters
with abo ut 1,900 ha . T hus, in only three yea rs, 60% of the
producers, rep resenting almost tw o-thirds of the area,
hav e a dopted the new technology. This p rovided an
exc ellent opportunit y to look a t potential benefits in terms
of pesticide use and impacts on health.
Pesticides in South Africa
In So uth Africa pesticides are used extensively a nd
intensively by small-sca le (mo stly black) fa rmers enga ged
in cash cropping (Dinham , 1993). Tab le 1 shows the main
insecticide s used in the Makhathini Flats, with
pyrethroids forming the bu lk of pesticide use prior to t he
introduction of
Bt co tton. S ince the u tilization of
pes ticides is the norm for pest control, South Africa
remains locked in a ‘pestic ide cu lture’ (Rother, 20 00).
Rother noted that the pe stic ide cu lture underplayed the
potential hazards of pesticides by ex plic itly and implicitly
rat ionalizing the support for pestic ide use. In South Afri ca
there remains widespread su pport for the use of
pes ticides and they are seen positively as a means of
maximizing profits a nd increasing rural household
eco nomy (Ro ther, 2000). However, London (199 8)
reported t hat acute pesticide poisoning accounted fo r 10%
of poison centre co nsul tations in South Africa, and w as an
imp ortant reason fo r admission to respiratory intensiv e
units. Furthermor e, Yousefi (20 00) noted a high incidenc e
of chemical poisoning (acute and chronic) among small-
holder farmers in So uth Africa. Rot her (2 000) further
pointed out that women and children were more at risk as
they d id most of the farm work and children would of ten
pla y in the field soon after a pesti cide application.
Methodology
The study is based on an in-dep th survey of 32 small-
sca le farmers carried out in January 2 002. The primary
aim of th e survey was to assess the impact of
Bt co tton on
pes ticide use, i nclu ding the dimension of fa rm labour
(Ismael
et al, 2002b).
A questionnaire was designed to ca pture the househo ld
structure of the farmers a nd their backgro und,
perceptions about
Bt cotton and the rationale for its
adoption, fa rmin g practices, their underst and ing of
pes ticides, and methods of disposal. It also captured
details of the different cotton plot sizes, their distance
from t he hou sehold dw elling, distance from the nearest
water source, ty pes of pesticide used, number of sprays
and time taken for ea ch spray, number of labourers and
other costs. In order to select a stra tified sample that
would be rep resentati ve of the Makha thin i region, the
Hlo koloko farmer as sociation was selected. There are 34 7
farmers in the Hlokolo ko farmer as sociation, and a
previous study (Ismael
et al, 2002b) showed that 60% were
fem ale, most farmers had farms of under 2 ha, an d a lmost
all farming transactions ( 99%) were based on credit.
Int erviews took approximately four hours, and were
combined with f ield visits and mea surement s of land
area, labour, etc.
There are three types of cotton f armer in the
Makhathini Flats: th ose who grow only
Bt cotton
(adopters), those who grow only non-Bt cotton (non-
adopters), and a third type, ‘balancers’, who grow both
Bt
cotton and non-Bt
cotton. The balancers are adopt ers of
Bt
cotton with a s mall proportion of non-Bt cotton (u sually
four hectares of
Bt an d o ne ha of non -Bt ). The sa mple
125
Bt cotton, pesticides, labour and health
Ou tlo ok on AGRICULTURE Vol 32, No 2
Table 1. Main insecticide s used and pes ts occurring in
Makhathini Flats, South Africa.
Trade name Acti ve in gredient Type Main target
Decis Deltam ethr in P Bollworm complex
Cyperm ethrin Cyperm ethrin P “ “
Ripcord Cyperm ethrin P “ “
Fen valerate Fenva lera te P “ “
Monostem Monocrotophos OP Aphids, jassids
Dim et Dim ethoate OP Thrips
Metasystox Dem eton -S-methyl OP Gra sshopper, thrip s
Rogor Dimethoate OP Aphids, jassids
Nuvacron Monocrotophos OP Aphids, jassids
Endosu lfan E ndosulfan O C Bollworm
Cruiser Thiamethoxam Nic otinoid Thrips, aph ids,
jassids
P = pyrethroid; OP = organophosphate; OC = organochlorine.
use d in the survey comprised 16 adopters, 8 balancers
and 8 non -adopters, nu mbers that reflect the preponder-
ance of these groups in the Hlokoloko farme r association
as a whole, based on cen trally held credit records. Other
cha racteristics of the sample were as follows:
Gender: The ratio was 43% ma le to 57% female. This is
typ ical of the region , where males tend t o work away
in the city.
Household characteristics: Women headed 60% of the
households in th e survey. On avera ge there were s even
members per household, and five full-time co tton
labourers (classified as those over 15 yea rs old ), three
of whom were female.
Main income: All o f the farmers surveyed had c otton
pro duction as their main income source, and 70% listed
maize as their second income source. Howeve r, ma ize
is not sold fo r cash but grown for household
consumption.
Farm characteristics: Most farmers (60%) owned between
one and five ha, with 30% having less than one ha and
10% having more than five ha.
Plot characteristics: There were 52 plo ts of cotton be-
lon ging to the 32 farmers in the sur vey, 63% of which
were Bt cotton and 37% non-Bt cotton. The plots were
1.7 ha on average, 1.1 km from the nearest w ater so urce
(pool o f water) and 1.2 k m from the farmer ’s dwelling.
Results
(a) Pesticide use
Only 37% of farmers relied o n t heir person al experience
whe n determining what pesticide to us e a nd whe n to
spr ay. Otherwise , a relative (1 9%), the c hairman (2 8%) or
an extension agent (1 6%) wo uld advise them. N ine pe r
cen t w ere
ver y satisfied with the advice given, a nd 91 %
were satis fied with the advice given. However, most of
the respondents mea sured the amounts of pesticide to be
use d by guesswork and o ften without the use of any
gra dua ted device. For convenience, the ma jority (62%) of
the respondents hid the pesticide in their fields during the
sprayin g sea son. A minority of respondents (38%) used a
purpose-built storeroom next to their house. F armers
would nor mally burn the chemical containers (4 4%) , but
mo re worrying ly, 25% reu sed the containers for carrying
water or to store other fluids. A lso of concern is that 31%
of the farmers surveyed t hrew the empty containers in
fields where children pla y a nd animals graze.
(b) Characteristics of Bt adopters
The adopters had a n average age o f 49 years, while for
non-adopters it wa s 45 yea rs. The adopters grew on
average four ha of cotton, com pared with six ha for the
balancers an d t hree ha f or the non-adopt ers. The
balancers ha d m ore cotton ex peri enc e (ave rage of six
yea rs) and the non-ado pters had the least experience
(average of three years). A dopters ha d a n average of four
yea rs’ experience growing cotton. These relativ ely short
periods reflec t in part the recent history of crop
pro duction in the M akhathini Fla ts and the vola tility of
cotton production in the a rea, wh ere so mu ch depends
upon the av aila bility of credit.
Over half (52%) of the adopters grew
Bt co tton becau se
the y conside red it to be labour-saving; 19% c hose
Bt
cotton because it needed less spraying; but only 19 %
adopted it because of yield benefit s. The balan cers
comprised some farmers who were growing
Bt in order to
compare it with the convent ional cotton so tha t they
could make their o wn judgment. Howev er, most o f the
balancers gro w some non-Bt in order not to become too
reliant on just one variety (ie as a form of in sura nce). A ll
the non-adopters were aware of
Bt co tton, b ut they d id
not use it bec ause of the high cost of seed (49%) and lac k
of finance (38%). The remaining 13 % d id not know why
the y did no t adopt it, or did not want to answer. A ll the
non-adopters said they would have grown
Bt co tton if
finance had been ava ilable to buy the more ex pensive
seed.
(c) Pest incidence
Farmers were asked about the in cide nce of bollworm on
their plots of cotton, and Figu re 1 presents th e results. All
but 7% o f the adopters a nd the balancers perceived a
dec rease in bollworm. Ho wever, 12.5% of non-adopters
saw an increase in bollworm, whilst 75% perceived no
change, and 12.5% a dec reased incidence. The re sults
suggest that
Bt co tton is effective in reducing the
bollworm population.
(d) Refugia
Given the high adopt ion rate of
Bt co tton in Makhathini
(th e latest estimate being 92%, based on seed sales figures
in 20 01
/
2002), the inclusion of refugia (plantings of crop
tha t are not
Bt-transformed or treated with
Bt) ha s been
seen as a crucial element in minimizing the risk o f insect
resi stance by both government agencies and c ommercial
companies. The results rev ealed that 5 0% o f t he
resp ondents were given some form of tra ining, usually by
Mo nsanto, and o f this pe rcentage some 56% had what
the y referred to as a refugium. However, only 10% under-
sto od the concept, and 70% of those given tra ining
believed that refugia were no t essentia l.
(e) Spray savings
Farmers were asked about the number of cotton sprays
the y used pre- and post-Bt ado ption. The re has been a
126 O ut loo k o n A GRICU LTURE Vol 3 2, N o 2
Bt cotton, pesticides , labour and health
Figure 1. Perception of bollworm in ciden ce among the three
types of farmers.
Note: Histograms indicate whether farmers’ perception s of
bollworm incidence have increased, decreased or remained the
same over the past three years. Farmers are divided into three
categories: Bt a dopters, non-a dopters and those who grow both
Bt and non-Bt varieties (balancers).
redu ction in the a vera ge number of sp rays per season
(from 11.2 to 3.8 sprays) for those who have adopt ed
Bt
cotton. All the farmers will still need to spray at least four
times per season for p ests not a ffected by
Bt resistance,
suc h as aphids, ja ssids, thrips and o cca sionally
grasshoppers. The reduction in the number of sprays is a
direct result of the reduct ion in pesticide used to control
bollworm. Cru iser (thiamexotham), a new insecticide t o
control thrips, aphids and jassids, is a powder mixed wi th
the seed before planting. This could further red uce the
number of sprays per season.
The reduction in the nu mbe r of sprays per season will
inevitably affect costs, amount of labour and distanc e
walked. Table 2 illustrates the savings th at could be made
per sp ray per hectare, and sa vings mad e p er hectare per
sea son. An average spra y per hectare will take 4.6 hou rs,
7.2 knapsack loads and 118.1 litres of water, and the
person spraying would walk over 911 metres. Water is an
exp ensive co mmo dity in th e M akhathini Fla ts and
children u sually collect it, often at a distance of over 1 km
from the field, and se ll at SAR5 per 25 litres. The
assump tion for the calculation in Table 2 is ba sed on the
amount of cypermethrin (for bol lworm control) used per
spray. Th e costs of other pesticides that all farmers would
Table 2. Costs of spraying one hectare of cotton with
cypermethrin for bollworm control in Makhathini.
Inp ut s In pu t s/ ha Co s t/ h a In pu ts /h a C os t / ha
for one for one for a tot al for 7
spray s pray of 7 spr ays s pray s
Kna psack (loads) 7 .2 – 5 0 –
Cypermethrin (m l) 42 3 31.73 2,961 2 22.11
Water (litres) 118.1 23.62 827 16 5.34
Time taken (hours) 4.6 11.50 32 80.5
Distance walked in
spra ying (m) 9,110.8 – 63,776 –
Total cost (SAR) – 66.85 – 4 67.95
Notes: Calculations are based on the assu mption that seven
sprays of cypermethrin are typically req uired per h ectare during
a cotton-growing season. The distance walked has been
calculated ba sed on the typical spraying m ethod emp loyed by
cotton farmers in the area. Spray costs exclude the cost of helpers
and depreciation of the knapsack sprayer. Wate r for spraying ha s
been calculated on the ba sis of SAR0.2/litre, and spra y labour
cost is calculated on the basis of S AR10 per four hours.
still need to spray are not included. Each a pplication of
cyp ermethrin wo uld co st t he farmer around SAR66 .85 per
ha. This figure excludes the costs of helpers used, as they
are usually household m emb ers. Th e spra y labour cost is
per man-hour, which cost SAR10 for four hours of
spraying. In any one season, a farmer would sa ve seven
sprays and this would translate into a sub stantial saving
of SAR467.98 (around $51) per season.
(f) Additional benefits
There are additional sa vings as a result of fewer pesticide
applications. The person spraying would no rmally walk
aro und 9.1 km per ha and, in order ensure a good
application o f pesticide, this could increase to 20 km in
lat er stages of crop growth and w ith higher bollworm
infestations.
This a mou nts to 63 km per ha per seaso n.
Spraying in the hot wea ther (around 37–40ºC ) is a
daunting task. None of the fa rmers in the survey used an y
pro tective equipment since they believed it was not
nec essary or wa s too uncomfortable.
Equal numbers of men and women do the spraying.
For every hecta re o f
Bt co tton grown, a farmer would save
two days of work (on e d ay spr aying and one d ay
collecting water). This saving is particula rly essenti al to
the women, g iven the many o ther house hold tasks in
whi ch they are involved. Most of the helpers a re children,
usu ally boys, who assi st with c olle cting wat er, carrying
spraying equipment, and disposin g of unused pestic ide.
Once again the time saved could potentially be o f benefit
for them.
(g) Health
Hospital adm ission rec ords in the Makhath ini area
covering the period from October 1997 to April 2001 were
che cked f or incidenc e of accidental insecticide poisoning
(marked a s ‘cotton’ in the records). Figure 2 sho ws the
tren d in ac cidental poisoning from 1997 to 2001 plotted
alo ngside the adoption rate of
Bt co tton. Farme rs begin to
grow cotton in November and harvest in May. The period
of inse cticide sp raying is between Decemb er and March .
127
Bt cotton, pesticides, labour and health
Ou tlo ok on AGRICULTURE Vol 32, No 2
Sin ce the introduction of
Bt
cotton in 1997 there has been
a rap id decline in co tton poisoning, which is probably
attributable in part t o the redu ction in pesticide required
for
Bt
cotton. Ho wev er, changes in the type of pesticide
may a lso have contributed. Dangerous organophospha te
insecticide s such a s rogor and endosulfan are no longer
use d to control bollworm, and use of their replacements,
the less toxic pyrethroi ds (cypermet hrin and
deltamethrin), ha s f alle n s ignif icantly with the
introduction of
Bt (Ismael
et al, 2002b). A local hospital
sup erintendent who had been wo rking in the region for
over 15 years confirmed that the requirem ent for atro pine,
use d as an antidote for org anophosphate insecticide
poisoning, ha d d eclined sharpl y sinc e
Bt cotton was
introduced in 1997 (Grant, 2002). Hence not a ll of the
redu ction in Figure 2 can be attributable to the
introduction of
Bt, and further researc h is required to
disentangle causes. It should also be noted tha t there is a
wid esp read under-notification of incidents of pesticide
poisoning in the area. This is par tly becaus e chronic
effects due to long-term exposure and some minor acute
effects that do not warrant admission, su ch as skin and
eye problems, are not reported.
(h) Effect of Bt on price variation
Ado pters and non-adopters we re asked at w hat pr ice pe r
kilogramme they wo uld stop grow ing cotton. The
adopters were more resilient and might stop at SAR2.10/
kg of co tton. H owe ver non-adopt ers wo uld stop grow ing
whe n the price fell to SAR2.40/kg of cotton. The reason
for this is tha t adopters have lower costs and higher
margins (Ismael
et al, 2001), t hus allowing them to absorb
grea ter pr ice variations.
Conclusion
The use of
Bt cotton appears to have some benefits for
farmers, workers and th eir fam ilies. The individua l
eco nomic b enef its resulting from less spraying can amount
to around SAR66 per spray per season, which tra nslates to
aro und S AR4 67 per ha per s eason. Savings come in the
form of lower pesticide and water u se, as well as lower
lab our c osts for spraying. However, there is t he added
benefit of less physical work whe reby a person spraying
would sa ve a round 9 km of walking per spray p er ha.
There are social implicatio ns tha t could arise from
these savings. The reduction in the n umb er of spr ays
would certainly free up women’s time to care for children,
or fo r income g eneration, and children could spend more
tim e studying or in other beneficial ac tivi ties. M ore
rese arch is ne eded t o examine how these savings might
link into other aspects of livelihood and leisure.
The higher the adoptio n r ate of Bt cotton, the lower th e
num ber of accidenta l in secticide poisonings. G iven the
findings on training, use, storage and disposal, the high
incidence of insecticide p oisoning (see also London, 1998 )
is hardly su rprising a nd more needs to be do ne to address
these limitations.
The resilience to price variation offered by
Bt
cotton
gives smallholder farmers some protection, since an
oversupply of world cotton (possibly even as a resu lt of
higher yields from
Bt
cotton) could drive do wn the world
market price.
Fig ure 2. In ciden ce of accidental insecticide poisoning and Bt
adoption rate for the Makhathini Flats between 1997
/
98 and
2000
/
01.
Note: Data are presented for the peak cotton-spraying months of
Dece mber to March.
One of the most worrying outcomes fro m t he survey
was the lack of farme r u nderstanding of refugia and their
poor pra ctice o f the t echnique (even when they were
aware of its importance), d espite t he extensive efforts of
Mo nsanto in promot ing refugia in Makha thini. This c ould
aff ect the longer-term susta inability of
Bt co tton, a lthough
the re is currently some resea rch looking a t the possible
imp act of alternativ e hosts.
Alt hough this c ase stu dy is based on a small sam ple, it
has highlig hted the po tent ial benefits of
Bt co tton. F urther
and possibly longer studies are requ ired to un derstand
fully the impact of this new te chnolog y on farmers.
Acknowledgme nts
We would like to thank Timothy Gumede and all the
farmers to whom we have talked during this survey.
Tha nks are also du e t o C harles Matlou and Wally Green
of Monsanto, So uth Africa. This research wa s funded by
the Department of Agricultura l a nd Food Econom ics,
University of Reading, UK.
128 O ut loo k o n A GRICU LTURE Vol 3 2, N o 2
Bt cotton, pesticides , labour and health
Referen ces
Benedict, J. H., and Altman, D. W. (2001), ‘Co mmercialis ation of
transgenic cotton expressing insecticidal crystal protein’, in J. J.
Jenkins and S. Saha, eds, Genetic Improvement of Cotton:
Emerging Tech nologies, Science Pu blishers, E nfield, pp 137–201.
Bennett, A. (200 2), ‘The impact of Bt-cotton on small holder
prod uction in the Makha thini Flats , South Africa’, Bt Cotton
Report, Website: http://www.mon santoafr ica.com.
Betz , F. S., Hammond, B. G., and Fuchs, R. L. (2000), ‘Safety and
advantages of Bacillus th urin giensis-protected plants to control
insect pests’, Regulatory Toxicology and Pharmacology, Vol 32, No
2, pp 15 6–173.
Dinh am, B. (1993), Pesticide Hazard. A Global Health and Environ-
ment A udit, Pesticide Trust, Zed Books, London.
Edg e, J. M., Benedict, J. H., Carroll, J. P., and R eding, H. K.
(2001), ‘B ollga rd cotton and assessm ent of global economic,
environmenta l and so cial benefits’, Journal of Cotton Science, Vol
5, pp 1–8.
Gra nt, A. (2002), Chief Medica l Officer, Bethesda Hospita l, South
Africa, personal communication.
Ismael, Y., Bennett, R., and Morse, S. (2001), ‘Ca n farmers in the
developing countries benefit from the modern technology?
Experience from Ma khathini Flats, The R epublic of S outh
Africa’, Crop Biotech Brief, Vol 1, No 5.
Ismael, Y., Bennett, R., and M orse, S. (2 002a), ‘Fa rm-level
impact of biotechno logy: smallhold er Bt cotton farmers
in South Africa’, Outlook on Agriculture, Vol 31, N o 2, pp 1 07–
111.
Ismael, Y., Bennett, R., and Morse, S. (2002 b), ‘Bt cotton and
pesticides. A case study of small holder farmers in Makhathini
Flats South Africa’, paper presented at the 6t h Intern ational
Conference on Agricultur al Biotechnology: New Avenues for
Prod uction, Consumption and Technology Tra nsfer, Ravello ,
Italy.
London, L. (1998), ‘Improving surveillance for acute pesticide
poisoning in the Western Cape, South Africa’, Website: http://
www.occuphealth. fi/e/info/anl/ 198/lond on.htm.
Manuna, F. T. (1 997), ‘Use of agricultu ral pesticides in Afr ica
with special reference to rural women’, African Newsletter, No
299.
Monsanto (2002), ‘B t cotton and product description’, available
online at http://www.monsanto.com/mons anto/ product.
Mullins, J., an d Mills J. (1 999 ), ‘Economics of Bollgard versus
non -Bollgard cotton in 19 98’, Proceedings of the Beltwide Cotton
Conference, Vol 2, pp 958–961.
Myers, D. (2000), ‘Cotton and pesticides’, New Internationalist, May.
Novillo, C., Soto, J., and C osta, J. (1999), ‘R esultados e n Esp ana
con vreidades de algodon protegidas geneticamente contra las
orugas de la s casuals’, Boletin de Sanidad Vegetal Plagas, Vol 25,
pp 383–393.
Pray, C., M a, D., Huang, J., a nd Qia o, F. (20 01), ‘Impact o f Bt
cotton in C hina ’, World Development, Vol 29, No 5, p p 81 3–825.
Rother, H. A. (2000), ‘Influences of pesticide risk perception on
the health of rural South Afr ican women a nd children’, African
Newsletter, No 2, Website: http://www.occuphealth.fi/e/ info/
anl/20 0/pesticide05.htm.
Stark, C. R . (19 97), ‘Econo mics of tran sgen ic cotton: some
indications based on Georgia producers’, Proceedings of the
Beltwide Cotton Conference, Vol 1, pp 251–253.
Traxler, G ., Godoy-Avila, S., Falck-Zepeda, J., and Espinoza-
Arellano, J. (2002), ‘ Transg enic cotton in M exico : economic and
environmental impacts’, Proceedings of the 5th Intern ation al
Conference on Biotechnology, Ravello, Italy. (Also in
Kalaitza ndonakes, N ., ed, Economic and Environmental Impacts of
First Generation Biotechnologies (in press).
Wier, A. T., Mullins, J. W., and Mills, J. M . (1998), ‘Bo llgard
cotton-update and economic comparisons including new
varieties’, Proceedings of th e Beltwide C otton Conference, pp 1 039–
1040.
Wilkins, T. A., Raja sekaran, K., and Anderso n, D. M. (2000),
‘Co tton biotechnology’, Critical Reviews in Plan t S cien ces, Vol
19, No 6, p 511 .
Xia, J. Y., Cui, J. J., Ma, L. H., Dong, S. X., and Cu i, X. F. (1999),
‘Th e role of transgenic Bt cotton in integrated insect pest
management’, Acta Gossypii Sinica, Vol 11, No 2, pp 57–64.
Yousefi, V. O. (2 000), ‘Agrochem icals in South Africa’, Website:
htt p://w ww.o ccu p hea lth .fi/ e/ in fo/ an l/1 99 /ag ro0 3.h tm.