Limits of Insecticide Cockroach Control in Council Flats in France

jem p115 13-11-95 09:40:50

Journal of Environmental Management (1995) 45, 379–393

Limits of Insecticide Cockroach Control in Council Flats in France

C. Rivault and A. Cloarec

Laboratoire d’Ethologie, CNRS URA 373, Universite de Rennes I,Campus de Beaulieu, 35042 RENNES cedex, France

Received 5 January 1995

The aim of our study was to analyse the results of controlled series ofprofessional insecticide treatments on feral cockroach populations in councilflats in three similar multi-family buildings, and to try to understand whycontrol success could be so low. Cockroach abundance and six generalenvironmental characteristics were recorded after each treatment for each flattreated. Cockroach abundance was divided into five levels. The environmentalcharacteristics recorded for each flat were: nationality of residents, duration ofoccupation, number of residents/m2, application of domestic pesticides,reactions of residents to the presence of cockroaches, and degree of cleanliness.

This series of treatments, proposed every two months, did not improve thesituation as greatly as expected. The presence or the absence of cockroaches inflats can be explained at least partly by some environmental factors like humandensity in the flat, building effect and cleanliness, but no single factor explainsthe size of the cockroach population.

The usual reasons evoked to explain pest control failures, such as cockroachinsecticide resistance, inefficiency of treatments due to formulation, proportionof constituent in compound, periodicity of application, etc., all influence theresult of a treatment, of course, but our data show that the part played by theresidents in a given dwelling has been greatly underestimated so far. When theresults of treatments were analysed, the results depended largely on theresidents’ attitude. 1995 Academic Press Limited

1. Introduction

The definition of urban pest species is applied to organisms considered damaging tohuman interests (Bennett and Owens, 1986; Hill, 1987) and is based both on subjectiveopinion as well as on objective analyses. Cockroaches are considered to be urban pestsfor many reasons. People dislike them or are afraid of them (Kellert, 1980); accordingto Wood et al. (1981) they are the animal least liked by American people.

Cockroaches rarely cause important economic damage, except possibly to electricalhousehold equipment or to food which can be spoilt by their contact. However,cockroaches are important from medical and public health points of view (Roth andWillis, 1957; Cornwell and Mendes, 1981; Cochran, 1982; Bennett and Owens, 1986;Cloarec et al., 1992; Rivault et al., 1993a). They present a contamination potential in

3790301–4797/95/040379+15 $12.00/0 1995 Academic Press Limited

Limits of cockroach control by insecticide380

food-handling facilities (Gorham, 1991; Rivault et al., 1993b) and hospitals (Burgess,1984).

Cockroach pest species are successful in exploiting the resources that humanstructures and life styles offer. Humans shape the urban ecosystem which providesall the necessary resources (food, water, harborage, temperature and humidity) forcockroaches, in an environment where they have no threats other than from humanbeings (Cornwell, 1968; Bennett and Owens, 1986; Rivault et al., 1994). Once acockroach population becomes well established, no improvement in sanitary conditionswill eradicate the population because of the adaptability of this pest (Bennett andOwens, 1986). The species Blattella germanica (L.) has a relatively short life cycle, greatreproductive potential, and often becomes widely distributed throughout a buildingafter it has reached a high population level in the initial infestation area (Rivault, 1992).

The objective of using insecticides is to reduce the pest infection to below the levelat which it causes damage, as effectively and as economically as possible (Dent, 1991).The ideal level of cockroaches is zero individuals in urban environments, makingcontrol or management of these organisms, described as pests, very important. However,pest control relying solely on pesticide use is often a failure for many reasons. Eradicationof cockroaches is a practical task involving skill and experience. Chances of successare higher if available information on this insect in its environment and on the propertiesof the insecticides are taken into account (Cromwell, 1976; Rivault, 1992). Wood (1980)suggested that professionals applying insecticide should be given time to do a thoroughjob and then to establish a monitoring schedule to ensure early detection and eliminationof any new infestations (Gorham, 1991).

One of the greatest problems facing the pest control industry and managers ofurban pest control programmes is that of economics. Pest elimination is admittedlyexpensive. Frequently, in order to secure business, agencies and companies will submitminimal bids and provide inadequate services (Bennett and Owens, 1986). Other reasonsfor the inefficiency of a control programme can be the lack of sanitation, structuralcomplexity, or lack of cooperation from the homeowner or client. Education of thepest control industry and of the public is another major problem confronting urbanintegrated pest management (Byrne et al., 1984). In the case of cockroach control inlow-cost public or private housing, the housing administrators have to make thedecisions about treatments (Gorham, 1991).

In the U.S.A., insecticide resistance is the major biological problem contributing todecreased efficiency of many conventional insecticides and pest management strategiesused for years by the industry. The presence and historical development of insecticideresistance in Blattella germanica is well documented, although data for the U.S.A. givevery variable pictures (Cochran, 1984, 1987, 1989, 1990, 1991; Bennett and Owens,1986; Schal and Hamilton, 1990; Vagn Jensen, 1990; Rust and Reierson, 1991; Rossand Cochran, 1992; Ross, 1993; Rust et al., 1993). According to Cochran (1991), pestcontrol operators treat monthly (sometimes more often) so that every generation islikely to be exposed at least once.

In France, pesticide use is generally limited to one professional treatment per yearand sometimes even less, because, before intervening, administrators wait for repeatedcomplaints from the residents who in the meantime try to manage with differentcommercial formulations like bombs, traps, baited food, etc., and who do so more orless successfully. Furthermore, as far as we know, no studies on cockroach resistanceto insecticides have been published for French populations: this, however, does notmean that resistance does not exist.

C. Rivault and A. Cloarec 381

T 1. Treatment frequency. Time from firsttreatment on day 1, in days

Treatment number Time (d)

T1 1T2 60T3 120T4 180T5 240T6 300T7 360

The aim of our study was to analyse the results of controlled series of professionalinsecticide treatment on feral cockroach populations in council flats in three similarmulti-family buildings and to try to understand why control success could be so low.

2. Material and methods

2.1.

Three blocks of council flats (named A, B and C) with high cockroach infestation levelswere chosen after preliminary investigations. Self-reported presence of cockroaches wasthen 45·7% for building A, 34·4% for building B and 54·3% for building C. These threebuildings were all built in 1972, in a residential area consisting of large blocks of flatsdeveloped by the city of Rennes (France) in the late 1960s and early 1970s. These 15-storey, 96-flat buildings all had the same lay-out: on each storey, there were six flatsopening onto the same landing with lifts at one end of the landing, stairs at the otherend and a common rubbish-shoot in the middle. The flats in any one building differedin size: on each storey, there were two one-bedroom flats, one two-bedroom flat, twothree-bedroom flats and one four-bedroom flat. The total number of flats differedslightly between buildings (A=94, B=92, C=96) because some of the ground floorflats had been converted for public use.

2.2.

A different type of insecticide, chosen from the three main groups of insecticidescommonly used, was applied in each building. Propoxur, a carbamate, was used inbuilding A. Building B was treated with deltamethrin, a synthetic pyrethroid. BuildingC was treated with propetamphos, an organophosphate. Formulations which could bedissolved in water were preferred for spraying the flats. Applications were done byprofessional pest control agents working for the city of Rennes under our control.

We proposed a free treatment every 2 months for a year, i.e. a total of seventreatments (Table 1). This frequency was chosen because of the duration of thedevelopmental cycle of Blattella germanica (at least 2 months) so that cockroacheswhich had escaped a treatment, or their immediate descendants, would not have hadtime to reproduce before the next treatment.

Before the first treatment, all the residents were informed that their building hadbeen chosen for a special cockroach control programme and that their cooperation


was requested. Our aim was to treat as many flats as possible during this first treatmentto get as precise a picture as possible of the cockroach situation at a given time (day1) and to evaluate as exactly as possible the quantity and distribution of cockroachesin each building and in each flat. According to the building, 71–82% of the flats in thebuilding were treated during this first treatment. During the first treatment, we aimedat treating the kitchen, the hall cupboard containing hot water pipes, the bathroomand the toilet in all the flats.

Before subsequent treatments, we asked, in each flat, if any of the residents hadseen any cockroaches since the previous treatment, and, if so, where, when and howmany had been seen, and we offered to treat the flat again. The residents were free toaccept or refuse any treatment. Insecticide was then applied with respect to what theresidents said or required. These treatments could include, in addition to the areastreated during the first visit, the hall, radiators in bedrooms, the living room and, forthe larger flats, a second cupboard. We elaborated a flexible protocol that aimed atreducing constraints on residents as much as possible.

3. Data collecting

Several environmental characteristics of each flat as well as the number of deadcockroaches following a treatment were recorded. Five different levels of cockroachabundance were noted:

Level 1: flat was treated, but we did not see any cockroaches.Level 2: flat was treated, but we only saw a few, i.e. less than 10, cockroaches

(developmental stage was not taken into account).Level 3: flat was treated, the cockroach population was well established (between

10 and 50 individuals).Level 4: flat was treated, cockroaches were pullulating (over 50 individuals).Level 5: flat was not treated.

General environmental characteristics recorded for each flat were (Table 2):• Building (A, B or C), storey and flat number.• Nationality of residents (N1=French, N2=foreigners).• Duration of occupation of flat (O1=0–2 years, O2=3–10 years, O3=>10 years

occupancy).• Number of residents/100 m2 in the flat (F1=1–3 people/100 m2, F2=4–6 people/

100 m2, F3=7–9 people/100 m2, F4=10–13 people/100 m2).• Self-reported application of domestic pesticides (I1=none, I2=cannot remember

the trade mark and casual use, I3=know the trade mark and frequent use).• Reactions of residents to the presence of cockroaches (P1=indifferent, P2=

frightened, horrified, disgusted, P3=no reply, P4=never seen or had any cockroaches,P5=cockroaches are dirty, unhygienic, P6=neighbours are responsible for their pres-ence, P7=they cause damage or prejudice).• Degree of cleanliness (S1=clean and tidy, S2=average, S3=dirty and cluttered).

4. Principal component analysis with respect to instrumental variables (PCAIV)

Multidimensional methods allow an important quantity of information to be analysedtogether in a way which yields richer indications than separate comparisons. Theyreveal the main tendencies of a complex set of data, which can then be classifiedand yield synthetic graphic representations (Benzecri et al., 1973; Grennacre, 1984).


T 2. Distribution of environmental factors

Environmental factor Percentage

BuildingA 31B 34C 33NationalityFrench (N1) 75Foreigners (N2) 24Duration of Occupation:0–2 years (O1) 333–10 years (O2) 45

11–20 years (O3) 21Human Density1–3 people/100 m2 (F1) 404–6 people/100 m2 (F2) 397–9 people/100 m2 (F3) 14

10–13 people/100 m2 (F4) 05Use of insecticideNone (I1) 49Unknown make (I2) 18Make specified (I3) 32Reactions to the presence of cockroachesIndifference (P1) 09Frightened, horrified, digusted (P2) 18No reply (P3) 24Never seen or had any (P4) 15Dirty, unhygienic (P5) 10Neighbours responsible (P6) 17Damage, prejudice, treatment smells bad (P7) 05CleanlinessClean (S1) 82Average (S2) 13Dirty (S3) 04

Multivariate principal component analysis with respect to instrumental variables(PCAIV) enables the study of important quantities of information (Lebreton et al.,1991). It is a multidimensional method as opposed to more usual statistical methodswhich only compare one or two variables at a time. PCAIV links one set of information(X1), to another set of information (X2) obtained on the same individual units as X1so that one set of factors is explained by the other. The reduction of X1 and X2 byfactorial methods improves the strength of the relationship as the reduction extractsthe largest sources of variation, which tends to be non-random (Lebreton et al., 1991).

This method is used in ecology to study species–environment relationships. We useda PCAIV to explain the presence or absence of cockroaches in relation to environmentalconditions that, here, were represented by characteristics of residents and of their flats,and to reveal a typology of the flats and their residents. The analyses were computedusing Biomeco software (Lebreton et al., 1990). The data for the PCAIV were presentedin two tables where the flats were described by two sets of variables: density ofcockroaches (data set X1) and environmental factors (data set X2). The lines of the


400

100

0Days

% 50

90

80

70

60

40

30

20

10

100 200 300

Figure 1. Proportion of flats (in percent) treated in each building, in relation to time (d). Μ=A; Β=B;Ε=C.

tables represented the 232 flats in the study, the columns of the first table representedthe data set (X1) for cockroach levels for each of the seven treatments, and the columnsof the second table presented the data set (X2) for the six environmental factorsrecorded.

5. Results

In the three buildings, 452 insecticide treatments were applied during seven sessionswith 232 flats treated at least once. Some flats, 50 in all, were never treated becausethose residents always refused a treatment and did not let us in. These flats weredistributed over the three buildings. although there were more untreated flats in buildingA (A=21; B=15; C=14). The 50 flats that had never had any of our insecticidetreatments were not included because we could not estimate the cockroach populationin those flats which we had never entered, nor could we evaluate the environmentalfactors relating to these flats.

For two of the buildings (B and C), the demand for treatment decreased rapidlywith time. However, 30% of the residents still asked for a treatment after a year’sintervention in building A (Figure 1).

5. General description: principal component analysis with respect to instrumentalvariables

The PCAIV relates the cockroach abundance data set (X1) (Figure 2a, b) to theenvironmental data set (X2) (Figure 3) on the same factorial plane defined by axes 1and 2. These two factorial axes extracted 48% of the total variability (total inertia ofthe cloud of data points).

The pertinence of the environmental factors can be measured by the ratio (trace)of inertias which indicates the weight of the environmental factors in explaining thepresence of cockroaches. The environmental factors defined in our data (set of dataX2) explain 15% of the cockroach abundance data set (X1).

The PCAIV graphical display of the cockroach abundance data set (X1) has a first


(a)4

4

2

2

2

1

2

11

1 1

1

3

3

3

3

2 32

4

32

5

355

5555 1

52

7

6 5 3

5

3

7

1 2 4

163

54 4

7

7

1

2

6 1 3 4 5

1

372

(b)

Figure 2. Cockroach abundance (data set X1). Projection of the cockroach abundance data on the planedefined by axes 1 (vertical) and 2 (horizontal) of PCAIV. This graph represents the centres of the data pointsfor the 5 levels of cockroach abundance for each treatment (5×7=35 points). Each of these data points isrepresented once on graph a and once on graph b. Figure 2(a) indicates the 5 cockroach abundance levels.Figure 2(b) indicates the treatments. The ovals added to Figure 2(a) delineate the three categories of flats

described in the text.

axis (ordinate) that explains the presence–absence of cockroaches; this axis separatesflats according to the level of cockroach abundance (Figure 2a). Data points forcockroach-free flats (abundance level 5) are situated toward the negative values of axis1, and the data for flats with cockroaches (abundance levels 1, 2, 3 and 4) are situatedtoward the positive values of this axis. There is a slight gradient on the graph fromlow cockroach densities (level 1) to high densities (level 4). No time effect was observedon this graph (Figure 2b). If the successive treatments had caused an important


P3

A

N2F3

I3

S2F4

P7

P1

S3

O3

O2F2 I2

P2

O1P6

P5

P4C

B

I1

N1 F1S1

Figure 3. Environmental factors (data set X2). Projection on plane defined by axes 1 (vertical) and 2(horizontal) of PCAIV of data points for the different categories of the 6 environmental factors recorded.

See Table 2 for details of abbreviations.

modification of the situation during this year’s study, data points for each treatmentwould have been near each other on the graph, and the centres of gravity for thedifferent treatments would form a succession along an axis, but this is not so; projectionsfor the different treatment points appear to be distributed randomly over the graph.The first axis can be interpreted as a cockroach presence–absence axis. The second axisis difficult to interpret in terms of cockroach abundance.

The second data set (X2) which includes the environmental factors indicates thathuman density, building and cleanliness contribute the most to the signification of axis1 (Table 3). The first axis that weights nearly equally these three factors appears to bea good compound variable to summarise variations in the environmental data set. Thisaxis opposes, for example, building A to the other two buildings, and very clean (S1)flats to medium clean (S2) and dirty flats (S3) (Figure 3). The cleanliness factor alsocontributes the most to explaining axis 2 (abscissa) and opposes different levels ofcleanliness (Figure 3). Axis 2 can be interpreted as a cleanliness axis.

The graphical representation of the PCAIV allowed us to separate the flats intothree categories (Figure 2a and Figure 3):

(1) The first category included clean (S1) flats with low human densities (F1) andwith French residents (N1), who did not use any insecticides (I1) and who said that


T 3. Contribution of the different environmental factorson the two main axes of the PCAIV for the three buildings

Factor Axis 1 Axis 2

Building 0·31 0·061Nationality 0·25 0·003Occupation 0·009 0·02Human density 0·40 0·13Use of insecticide 0·14 0·04Reaction to cockroaches 0·25 0·17Cleanliness 0·28 0·51

T 4. (a) Contribution of the different environmentalfactors on the two main axes of the PCAIV for building A


Nationality 0·086 0·020Occupation 0·026 0·008Human density 0·153 0·062Use of insecticide 0·078 0·013Reaction to cockroaches 0·190 0·288Cleanliness 0·632 0·605

(b) Contribution of the different environmental factors on thetwo main axes of the PCAIV for buildings B and C


Building 0·001 0·000Nationality 0·363 0·003Occupation 0·077 0·006Human density 0·654 0·332Use of insecticide 0·189 0·048Reaction to cockroaches 0·325 0·157Cleanliness 0·175 0·256

they had never had or seen any cockroaches in their flat (P4). Sure enough, we neverfound any cockroaches in these flats during any of our inspections (cockroach abundancelevel 5).

(2) The second category of flats included flats of average cleanliness (S2) with highhuman densities (F3 and F4), often with foreign residents (N2) who used insecticides(I3) and who either said that they were horrified or frightened of cockroaches (P2) orcomplained about the damage caused (P7). We recorded low or medium densities ofcockroaches (levels 1, 2 and 3) in these flats.

(3) The third category of flats included dirty and cluttered flats (S3) with highhuman densities (F3) and the presence of cockroaches left the residents indifferent (P1).Nationality was not an important factor. We recorded high cockroach densities in theseflats (level 4). However, there were very few flats of this category.


5.2.

After the general description of our data, we analysed each environmental factor inturn in relation to the cockroach density levels and their contribution to the two firstaxes of the PCAIV (Table 3, Figures 2, 3).

5.2.1. Duration of occupation

This factor was not related to the presence of cockroaches. The data points for thethree categories of occupation duration (O1, O2 and O3) were all projected near thecentre of axes 1 and 2 (see contribution of different factors to the two first axes inTable 3). The distribution of cockroaches cannot therefore be explained by changes intenancies. This does not support the wide spread idea that cockroach infestations aredue to moving home frequently.

5.2.2. Use of insecticide

This factor has a low contribution to axis 1 (Table 3). I1 (no use of insecticide) isassociated with flats without cockroaches and I3 is associated with flats with cockroaches.Residents in flats with few cockroaches only used insecticides occasionally and theydid not know the name of the insecticide they used, nor could they show us anyinsecticide formulations. The use of domestic insecticides implies the presence ofcockroaches, even though this use has little or no effect on the density of cockroachpopulations. On the other hand, when there are no cockroaches, residents did not usecockroach insecticides, or only very rarely.

5.2.3. Human density

This factor contributes the most to the first axis (Table 3). The link between humandensity and level of cockroach populations can be explained by the fact that there weremany objects (furniture as well as smaller objects) on the floors and on or against wallsin flats with high human densities. An increase in the amount of objects in the flat,thereby decreasing free floor-space and wall-space, plays an important part in thecontrol of cockroach populations by providing an increased number of available sheltersfor cockroaches and by impairing the efficiency of professional treatments, because itis often impossible to remove all the objects so that the insecticide can be appliedcorrectly everywhere.

5.2.4. Nationality

This factor also contributes to axis 1. This factor is strongly linked to human densitybecause French families are usually smaller than foreign ones. Small flats were occupiedby single people of French nationality. On average, low human densities are linkedwith French tenants and high human densities with foreign families. But the few largeFrench families in these buildings also had high cockroach densities.

5.2.5. Reactions of residents to the presence of cockroaches in their flat

Indifference to the presence of cockroaches (P4) was associated with high cockroach


densities (level 4). Fright (P2) and prejudice (P7) were associated with low cockroachdensities. The reply ‘‘never seen or had any cockroaches in their flat’’ (P4) was associatedwith flats without cockroaches. The projection points for the reactions ‘‘cockroachesraise hygiene problems’’ (P5) or ‘‘neighbours are responsible for their presence’’ (P6)are near the centre of axes 1 and 2. These two parameters were not related to thepresence–absence of cockroaches, but reflect more the prevailing opinion.

5.2.6. Cleanliness

This is the factor that contributes the most to axis 2 (correlation=0·51, Table 3) andopposes very dirty flats (S3) to average (S2) and clean (S1) flats. Axis 2 is interpretedas a cleanliness axis. The fact that axis 2 is at right angles to axis 1, interpreted as acockroach presence–absence axis, indicates that these two factors are independent. Inother words, the presence–absence of cockroaches was not related to cleanliness. Adirty flat will not necessarily harbour cockroaches and a very clean flat will not alwaysbe without any cockroaches. However, medium scores of cockroach densities (level 3)are linked with average cleanliness scores, and high cockroach densities (level 4) arelinked with scores for dirty flats (S3). The absence of cleanliness encourages cockroachpullulation but plays no part in their presence–absence.

5.2.7. Building

Building A was opposed to the other two buildings on axis 1. Building A had a higherlevel of cockroach infestation than the other two buildings. This confirms the indicationsfrom the analysis of the demands for treatment (Figure 1). As building A was opposedto the other two buildings on the plane defined by the first two axes of the PCAIV, thedata were subdivided and two more PCAIV were computed, one for building A dataalone, the other for data for buildings B and C.

Axes 1 and 2 of the PCAIV on building A data extracted 40% of the total variability.The environmental factors (X2 data set) explained 29% of the records of cockroachpresence (X1 data set). The factor ‘‘cleanliness’’ is the factor which made the highestcontribution to the two first axes (Table 4), axis 1 opposes S1 to S2 and to S3, andaxis opposes S2 to S3.

Axes 1 and 2 of the PCAIV on the data for the other two buildings extracted 52%of the total variability. The environmental factors (X2 data set) explained 21% of therecords of cockroach presence (X1 data set). For these buildings, human density is thefactor which made the highest contribution to the two first axes (Table 4).

After subdividing the data, 29% (for building A) and 21% (for buildings B and C)of the total inertia was explained by the environmental variables (X2 data set), whereasbefore separating the data for the different buildings these factors explained only 15%of the inertia. This is because of interactions between factors and because differentfactors dominate in the different subsets: cleanliness in building A and human densityin buildings B and C data. Nevertheless, the distribution of these factors did not differsignificantly between buildings (chi square, P>0·05 for both factors; Table 5). Thedifferences revealed by the PCAIV were not related to differences in distribution offactors F (human density) or S (cleanliness), but due to the fact that these factors arereally explanatory parameters for the observed differences in cockroach distribution.


T 5. Comparison of the distribution of human densityand cleanliness between buildings (number of flats in each

category)

Building F1 F2 F3 F4 S1 S2 S3

A 27 30 12 4 58 11 4B and C 67 61 22 8 134 19 5

See text for explanation of abbreviations (page 382).

6. Discussion

This one-year series of treatments, proposed every two months, did not improve thesituation as greatly as expected. The presence of cockroaches was still reported in ahigh proportion of flats at the end of these treatments in building A.

The multivariate PCAIV analysis of our data indicated that the presence or absenceof cockroaches in flats can be explained at least partly by some environmental factorslike human density, building effect and cleanliness. But none of the environmentalfactors which we recorded appear to be able to explain, alone, the presence ofcockroaches, or the size of the cockroach populations. This problem is complex andundoubtly multifactorial.

Human density is an indirect way of measuring free ground and free wall space ina flat, particularly in the kitchen where cockroaches are generally found. Even thoughthere is a good relationship between the number of residents/m2 and space overcrowdedwith furniture and various other objects, overcrowded flats with low human densitiesdo exist. Cockroaches find more potential undisturbed shelters in cluttered flats. Themore shelters there are, the more widely cockroaches are distributed all over the flat(Rivault, 1989, 1990).

The absence of cleanliness encourages the development of cockroach populationsbecause food resources become more dispersed throughout the flat. Shorter distancesbetween shelter and food make access to food easier and facilitate survival and growthof young larvae (Rivault and Cloarec 1990, 1991) and female reproduction (Cochran,1983).

The observed building effect is not tied to the building itself because the threebuildings were built according to the same architectural plan, during the same period,with the same types of materials. Building A differed from buildings B and C becausethe rent scale was lower, due to the particular financing of this building. Therefore,building A housed people with lower incomes than those in buildings B or C. Thediscrepancies between treatment results for building A and for buildings B and C canbe explained in part by differences in residents’ income and social class. Furthermore,the contribution of the two main environmental factors, cleanliness and human density,to the presence of cockroaches differed between buildings. The cleanliness factorcontributed the most to the presence or the absence of cockroaches in building A andthe human density factor contributed the most in buildings B and C.

After one year, some flats with low cleanliness levels and high human densities stillhad high cockroach densities that we had been unable to reduce by applying ourprotocol. This result can be explained by several reasons. The residents were asked toevaluate the presence of cockroaches in their flats. According to their feelings and theirupbringing, they report the presence of cockroaches for variable real cockroach densities:


other people are not very observant and only remark on the presence of these pestswhen there are many. In addition, residents are free to accept or refuse treatment, inaccordance with the French law of private liberty, and we found very often that accessto these flats was refused not only to us and the pest control agents, but also to theiradministration officials. In spite of all our attempts, we had great difficulty in contactingall the tenants. In the three buildings, the proportion of flats we had never been ableto visit was still relatively high at the end of the year (50/282, approximately 17%).Sawyer and Casagrande (1983) already stressed the importance of human factors inthe control of pest populations and stated that responses to arthropods can beconditioned by many factors such as cultural and educational background, availableinformation, personal sensibility, preferences and values.

When cockroach densities reach very high levels, highly infested flats play the partof a population reservoir for neighbouring flats; they become a source of migrantanimals especially when maximum capacity of the initial occupation zone has beenreached (Rivault, 1992). The available insecticides would be efficient for reducing theselevel 4 density cockroach populations to at least level 2 or 3 densities, as long as theseflats could be treated.

Rules stated by Wood (1980) for getting better results in control of urban pests areinteresting but only vaguely hint at the underlying social problems. His first rule is‘‘control of German cockroaches cannot be obtained without tenant cooperation’’. Howcan we get all the families to cooperate in a given building at the same time? Thesecond rule is ‘‘Pesticides are never a substitute for basic sanitation’’. This rule is onlypartly true. Improved sanitation in the few flats with important cockroach populations,which are dirty, cluttered, with high human densities and with residents who areindifferent to the presence of cockroaches (our third category of flats) should reducecockroach density. However, the majority of flats with cockroaches harbour smallcockroach populations and are clean with high human densities and with residents whotry to control these pests (our second category). Improved sanitation in these flats willnot eradicate the cockroaches. When a cockroach population is well established,increased cleanliness will not modify the fact that there are cockroaches in the flat(Bennett and Owens, 1986).

The usual reasons evoked to explain pest control failures such as cockroachinsecticide resistance, inefficiency of treatments due to formulation, proportion ofconstituent in compound, periodicity of application, etc., all influence the result of atreatment, but our data show that the part played by the residents in a given dwellinghas been greatly underestimated so far when the results of treatments are analysed.Here, we show that, under similar treatment conditions such as we applied in ourprotocol, the results depend largely on the residents’ attitude in each flat.

We are extremely grateful to all who facilitated this study; the Departement d’Hygiene et deSante de la Ville de Rennes and its pest control team, the Office des H.L.M. de la Ville de Rennesand members of these staffs with a special mention for M. Derennes, residents of buildings A,B and C who gave us free access to their flat. Roussel-Uclaf-Procida provided the Deltamethrinand Bayer (Department of Hygiene) provided the Propoxur. This study was supported by a grantfrom the Municipalite de Rennes and one from the Ministere de l’Environnement (DRAEI no

93070). R. Quris helped us compute the PCAIV.

References

Bennett, G. W. and Owen, J. M. (Eds) (1986). Advances in urban pest management. New York: Van NostrandReinhold Company. 398 pp.


Benzecri, J. P. et al. (1973). L’analyse des donnees. II. L’analyse des correspondances. Paris: Dunod. 619 pp.Burgess, N. R. H. (1984). Hospital design and cockroach control. Transactions of the Royal Society of

Tropical Medicine and Hygiene 78, 293–294.Byrne, D. N., Carpenter, E. H., Thoms, E. M. and Cotty, S. T. (1984). Public attitudes toward urban

arthropods. Bulletin of the Entomological Society of America 30, 40–44.Cloarec, A., Rivault, C., Fontaine, F. and Le Guyader, A. (1992). Cockroaches as carriers of bacteria in

multi-family dwellings. Epidemiology and Infection 109, 483–490.Cochran, D. G. (1982). Cockroaches—biology and control. Who, vector biology and control series. Geneva,

Switzerland 856, 1–53.Cochran, D. G. (1983). Food and water consumption during the reproductive cycle of female German

cockroaches. Entomologia experimentalis et applicata 34, 51–57.Cochran, D. G. (1984). Insecticide resistance in cockroaches: is it at a crossroad? Pest management 3, 26–31.Cochran, D. G. (1987). Selection for pyrethroid resistance in the German cockroach (Dictyoptera: Blattellidae).

Journal of Economic Entomology 80, 1117–1121.Cochran, D. G. (1982). Monitoring for insecticide resistance in field-collected strains of the German cockroach

(Dictyoptera: Blattellidae). Journal of Economic Entomology 82, 336–341.Cochran, D. G. (1990). Managing resistance in the German cockroach. Pest Control Technology 18, 56–57.Cochran, D. G. (1991). Extended selections for pyrethroid resistance in the German cockroach (Dictyoptera:

Blattellidae). Journal of Economic Entomology 84, 1412–1416.Cornwell, P. B. (ed.) (1968). The cockroach. Volume I. A laboratory insect and an industrial pest. London:

Hutchinson and Co. 387 pp.Cornwell, P. B. (ed.) (1976). The cockroach. Volume II. Insecticides and cockroach control. London: Associated

Business Programmes Ltd. 557 pp.Cornwell, P. B. and Mendes, M. F. (1981). Disease organisms carried by oriental cockroaches in relation to

acceptable standards of hygiene. International Pest Control 23, 72–74.Dent, D. (ed.) (1991). Insect pest management. Wallingford: C.A.B. International. 604 pp.Gorham, F. R. (ed.) (1991). Ecology and management of food-industry pests. Association of Official Analytical

Chemists, Arlington. 595 pp.Greenacre, M. J. (ed.) (1984). Theory and applications of correspondence analysis. London/Orlando: Academic

Press. 357 pp.Hill, D. S. (ed.) (1987). Agricultural insect pests of temperate regions and their control. Cambridge: Cambridge

University Press. 624 pp.Kellert, S. R. (1980). American attitudes toward and knowledge of animals: an update. International Journal

of the Study of Animal Problems 1, 87–119.Lebreton, J. D., Roux, M., Bacou, A. M. and Banco, G. (1990). Biomeco (Biometrie-Ecologie), version 3·9,

software of statistical Ecology for PC.Lebreton, J. D., Sabatier, R., Banco, G. and Bacou, A. M. (1991). Principal component and correspondence

analyses with respect to instrumental variables: An overview of their role in studies of structure—activityand species—environment relationships. Applied multivariate analysis in SAR and environmental studies.(J. Devillers and W. Karcher, eds) Luxembourg: ECSC, EEC, EAEC. pp. 85–114.

Rivault, C. (1989). Spatial distribution of the cockroach, Blattella germanica, in a swimming-bath facility.Entomologia experimentalis et applicata 53, 247–255.

Rivault, C. (1990). Distribution dynamics of Blattella germanica in a closed urban environment. Entomologiaexperimentalis et applicata 57, 85–91.

Rivault, C. (1992). Invasion des milieux urbains par les blattes: exemple de la ville de Rennes. In Actes ducolloque national d’ecologie urbaine, Mions (Rhone) (J. M. Legay, ed.) pp. 145–156. 27–28 septembre1991, Universite Claude Bernard, Lyon.

Rivault, C. and Cloarec, A. (1990). Food stealing in cockroaches. Journal of Ethology 8, 53–59.Rivault, C. and Cloarec, A. (1991). Exploitation of food resources by the cockroach Blattella germanica in

an urban habitat. Entomologia experimentalis et applicata 61, 149–158.Rivault, C., Cloarec, A. and Le Guyader, A. (1993a). Bacterial load of cockroaches in relation to urban

environment. Epidemiology and Infection 110, 317–325.Rivault, C., Cloarec, A. and Le Guyader, A. (1993b). Bacterial contamination of food by cockroaches.

Journal of Environmental Health 55, 21–22.Rivault, C., Cloarec, A. Mathieu, N. and Blanc, N. (1994). Les blattes, un animal de ville a toutes saisons.

Les Annales de la Recherche Urbaine 61, 33–38.Ross, M. H. (1993). Comparisons between the response of German cockroach field-collected strains

(Dictyoptera: Blattellidae) to vapor and contact with a cyflathrin formulation. Journal of EntomologicalScience 28, 168–174.

Ross, M. H. and Cochran, D. G. (1992). Strain differences in the response of German cockroaches(Dictyoptera: Blattellidae) to emulsifiable concentrates. Journal of Economic Entomology 85, 1201–1208.

Roth, L. M. and Willis, E. R. (1957). The medical and veterinary importance of cockroaches. SmithsonianMiscellaneous Collection 134, 1–147.

Rust, M. K. and Reierson, D. A. (1991). Chlorpyrifos resistance in German cockroach (Dictyoptera:Blattellidae) from restaurants. Journal of Economic Entomology 84, 736–740.


Rust, M. K., Reierson, D. A. and Zeichner, B. (1993). Relationship between insecticide resistance andperformance in choice tests of field-collected German cockroaches (Dictyoptera: Blattellidae). Journal ofEconomic Entomology 86, 1124–1130.

Sawyer, A. J. and Casagrande, R. A. (1983). Urban pest management: a conceptual framework. UrbanEcology 7, 145–157.

Schal, C. and Hamilton, R. L. (1990). Integrated suppression of synantropic cockroaches. Annual Review ofEntomology 35, 521–551.

Vagn Jensen, K. M. (1990). Cockroaches. In Annual report 1990, Danish pest infestation laboratory, Lyngby,pp. 52–57.

Wood, F. E. (1980). Cockroach control in public housing. Annales Entomologici Fennici 48, 14–18.Wood, F. E., Robinson, W. H., Kraft, S. K. and Zungoli, P. A. (1981). Survey of attitudes and knowledges

of public housing residents toward cockroaches. Bulletin of the Entomological Society of America 27,9–13.

Documents

Limits of Insecticide Cockroach Control in Council Flats in France