First published in Antenna, 21(1), 14-18. January, 1997
Scouting in cotton IPM - a cautionary note and a peculiar observation
Recently, while talking of the desperate
shortage of funding for entomological research (or, apparently, any
biological research other than that with DNA in the project title), a
distinguished Fellow commented that there was a perception that
"applied entomology" had failed. Now I suspect that the "failure" owes
more to the responses of politicians and economists (who, with rural
sociologists, now dominate most "Aid" projects) than to the technology.
The classic example, of course, is the infamous collapse of the Malaria
Eradication Programme in Sri Lanka. In the early 1960's, this went from
near-success to total failure principally because funding was cut at
the crucial point - there were so few cases that the high cost of
surveillance and insecticide-spraying was perceived as unjustifiable.
However, we have to take some of the blame, if only for professional
naiveté, and I proffer the following illustration.
Scouting is an essential element of integrated pest management (IPM)
programmes in many cotton growing countries and has been reviewed
recently by Matthews (1996). A feature of the history of insecticide
use for crop protection in cotton is that applications were directed
first against a single pest and with only a few applications per
season. An example of this is the perceived need to control the cotton
jassid, Empoasca lybica,
in the Sudan. This pest was regarded as the major problem in the 1950's
and was satisfactorily controlled with a single annual application of
DDT in mid-October. By the mid-1960's, however, several previously
minor pests had assumed major importance, especially the whitefly, Bemisia tabaci, and to a lesser extent, Helicoverpa armigera, and the spiny bollworm, Earias insulana.
The rise of these pests is suspected to have been due to elimination of
beneficial insects. By 1978 applications of pesticide in the Gezira
area were up to 14 per season, with aerial spraying of a range of
chemicals. Such levels were at, if not exceeding, the level at which
the crop became economically unviable (personal field notes).
A similar picture emerged in the Americas, where attempts to control the bollweevil, Anthonomus grandis, may have contributed to the rise in importance of the bollworms, notably Heliothis zea.
Matthews (1996) has suggested that excessive use of insecticide with
numerous applications per season came about because of the erroneous
belief by some farmers that more applications could mean greater yield.
Writing from experience in Central America, however, Schmutterer (1977)
described how the need "to combat the strength of the development of
destructive insects" and the need to protect the yield had led to 30-40
applications of pesticide per cotton-growing season. That was the
situation in October 1977, when I went to El Salvador to conduct a
series of field trials to evaluate the efficacy of amitraz, of
a then relatively unfamiliar type of insecticide, a formamidine, and to
investigate problems encountered with the initial commercial launch of
the product. What follows is essentially a description of those trials
but, out of respect to the commercial companies involved and as
technical details are not crucial to the theme, information on dose
rate, etc., is not given.
By the time of my trials scouting had been introduced and Schmutterer
(1977) described the criteria developed in Nicaragua and El Salvador
for H. zea
control. The target had to be the early larval instars (L1-2), as even
insecticides with a high toxicity were ineffective against larvae of
more than 1 cm in length. The fact that such larvae would already be
feeding inside the developing bolls was not mentioned but that alone
would protect them from the pesticide. In Nicaragua the curve for
economic damage had been calculated as showing that, in plantations of
more than 100 days since sowing, the first application of pesticide was
required when L1-2 exceeded 5,700/ha, and 7,000/ha thereafter. The
scouting technique in El Salvador was to inspect 2 m lengths of cotton
row for L1-2 larvae with a criterion of 1.2-1.5 larvae up to 60 days
from planting, 1.0-1.2 larvae from 70 days to the first picking of the
cotton fibre, 0.8-1.0 larvae from then until the second picking, and
1.0-1.2 larvae thereafter (Schmutterer, 1977). Many farmers were using
a modified system in which 25 or 50 plants were inspected in each
field. With plants being thinned to 50 cm apart, this was equivalent to
respective economic levels of 24-30 L1-2 larvae, 20-24 larvae, 15-20
larvae and 20-24 larvae per 100 plants (often expressed as a %)
The commercial introduction of amitraz
to control lepidoptera was complicated somewhat by it being thought to
be effective solely against the egg stage. By a field study of labelled
eggs, I was able to establish that the real effect was to kill newly
emerged or emerging L1 larvae, i.e. it was a neonaticide. Nevertheless,
amitraz had to be used in a mixed application with a pesticide
which would kill older L1-2 larvae. The very first applications in my
large scale field trials, with aerial ULV spraying, were successful,
with amitraz + parathion + parathion-methyl mixtures being more effective than parathion alone. Importantly the hatching success of H. zea
eggs was halved. The results led the host farmer to enthuse to all his
neighbours but when a second application appeared less successful he
became quite agitated. Further trials, on a better-run plantation and
with two new formulations of amitraz gave more consistent and
more encouraging results but there still seemed to be inconsistencies
when it came to reducing the L1-2 levels below the threshold. However,
those inconsistencies applied equally to the "standard" treatments (see
below) and I decided that it was necessary to investigate further.
The wider situation
During my first two weeks in El Salvador, I visited plantations in
three of the major cotton areas and was given information that some
commercial applications of amitraz
had not given satisfactory results. Because of the astuteness of many
of the plantation owners, some being large commercial enterprises and
very challenging over the potential price of the product, the results
of all pesticide applications were keenly scrutinised. Discussions
revealed that on 4th and 5th October 1977 several insecticides
(including profenofos, permethrin and parathion + parathion-methyl) had given good control of H. zea
but one week later, 11th October, those compounds were not performing
as well. With the introduction of a scouting system, the plantations
mostly kept very good records of all farm activities, Thus, I was able
to examine the data from insect monitoring (which included eggs, L1-2
and L3-4 stages, etc.) and make graphical analyses of the H. zea levels throughout September, October and November for three fields on a well-run plantation.
Figure 1. Results of farmers' applications
Data (shown as nos. per 100 plants) gathered by scouts at two
groups of fields at the Hacienda Sta. Lucia Orcoyo, El Salvador, 1978.
Days shown without data are when spraying took place or days on which
no assessments took place. Spraying at Lote La Isla took place on 24/09
(parathion + methamidophos), 29/09 (fenvalerate + parathion-methyl), 03/10 & 06/10 (parathion-methyl), 08/10 (parathion + parathion-methyl), 13/10 (monocrotophos), 19/10 (fenvalerate + parathion-methyl), 25/10 (parathion + parathion-methyl) and 29/10 (fenvalerate + parathion-methyl). Spraying at Lote Vainilla took place on 01/10 (fenvalerate + parathion-methyl), 04/10 (parathion + parathion-methyl), 08/10 (parathion-methyl), 12/10 (fenvalerate + methamidophos) and 17/10 (monocrotophos); after a 10 day break, the observations continued on Lote Orcoyo, with spraying on 29/10 (fenvalerate + chlordimeform) and 05/11 (monocrotophos)
From a simple eye-fit, the following points seemed to be of significance:
The apparent effect of an insecticide, in almost all instances, was
that it gave satisfactory control, from the point of view of the
farmer, or farm manager, when applied during a downward trend of the
oviposition macrocycle but gave unsatisfactory control when applied
during an upward trend. I felt that this perception was due primarily
to the dependency solely on the criterion of maintaining an H. zea population level of less than 25 L1-2 larvae per 100 plants. Examples of the end result are:
- There were macrocycles in the population levels of H. zea
with three or four maxima of oviposition in the three month period.
These appeared to be at 16-20 day intervals and were not influenced by
the phases of the moon in the manner reported by Nemec (1971) from the
- Underlying the macrocycle was a short cycle of oviposition with
peaks every four days. I found no other reports of such curious cyclic
behaviour but noted that Nemec's (1971) data showed short-term
fluctuations in oviposition which he attributed to environmental
- Paralleling the four-day oviposition cycle was a cycle of L1-2
larval levels indicating a four day incubation period and a four day
first instar period.
- Fenvalerate applied on 19th October reduced a white
egg level of 42 (on 18th October) to an L1-2 level of 22 (on 22nd
October) - a "successful" application despite a larval mortality of
only 47.6%. The same pyrethroid applied on 7th November reduced a white
egg level of 290 (on 6th November) to an L1-2 level of 80 (on 13th
November) - an "unsuccessful" application despite a mortality of 72.4%.
- Amitraz plus parathion on 10th October reduced a
white egg level of 84 (on 9th October) to an L1-2 level of 10 - a
"successful" application with a mortality of 88.1%. The same mixture
applied on 6th November reduced a white egg level of 304 (on 5th
November) to an L1-2 level of 58 (on 9th November) - an "unsuccessful"
application despite a mortality of 80.9%.
- Both "successful" applications were made (unknowingly) in a period
when the population macrocycle was in a downward phase and both
"unsuccessful" applications were made during an upward phase.
Clearly, the scouting concept was eminently sensible and aimed at
reducing the usage of pesticides. The evidence in 1977-78, when I was
involved in the development trials with amitraz,
was that it was not having the desired effect. Rightly or wrongly, the
consequences of the reliance on a simple criterion were to induce a
feeling of near-perpetual panic in all the farmers. On one occasion,
actually in Colombia where similar economic levels were being used, we
had to abandon a field trial solely because of this panic. My colleague
had vehicle problems and was unable to reach the farm at the planned
time of spraying (in the time-window between dawn when the crop has too
much dew on it and mid-morning when rising temperatures cause wind
levels to rise too high). The farmer ordered the trial site to be
sprayed with his normal pesticide, thus ruining our trial. On a wider
scale, also in Colombia, the early season release of parasitoid wasps, Trichogramma
sp., was proving quite effective and could be observed in the field as
obviously parasitised eggs. The moment an immigration of H. virescens
led to a rise in egg levels and L1-2 larval numbers exceeded the
threshold the farmers panicked and used a broad-spectrum pesticide,
wiping out the Trichogramma.
To return to the early problems with the commercial introduction of amitraz,
my subsequent investigations revealed that the dosage rates were
derived from "successful" trials by a well-respected government testing
station in Nicaragua. The parts of the official report containing
results with amitraz had been translated from the original
Spanish and these showed comparable efficacy with conventional
commercially-available products. Full reading of the report, however,
revealed that the overall infestation levels of H. zea in that
year had been relatively low and to reduce L1-2 levels below the
economic threshold had posed no problems with any pesticide. An
additional complication encountered during the pre-marketing
development stage with amitraz was that official testing
procedures, for instance in Colombia and Turkey, used larvicidal
efficacy as the criterion for registration for commercial use and the
authorities were unwilling or unable to accept the need to assess the
effect of the chemical on eggs and hatching.
To conclude, although the way forward has to lie in good IPM, it is
essential for the widest possible range of information to be gathered
before making recommendations and then a careful extension programme to
fully inform even well-educated and sophisticated farmers is vital.
Matthews, G.A. (1996) The importance of scouting in cotton IPM. Crop Protection, 15, 369-374.
Nemec, S.J. (1971) Effects of lunar phases on light-trap collections and populations of bollworm moths. Journal of Economic Entomology, 64, 860-864.
Schmutterer, H. (1977) Plagas y enfermedades del algodon en Centroamerica. Deutschen Gesellschaft für Technische Zusammenarbeit (GTZ) Gmbh, Eschborn, 95 pp.