Reprinted from ENVIRONMENTAL ENTOMOLOGY, Vol.. 12, No. 2. APKIL 1983

Factors Affecting the Relative Abundance of on Nectaried and Nectariless Cotton’

I. K. ADJEI-MAAFO’ ASD L. T. WILSON‘

Envirun. Entoniol. 12: 3-19-352 ( 1983) ABSTRACT This study further confirms that nectariless cultivar5 of cotton are less attractive than the near-isogenic nectaried cultivars when grown in small plots. The results also indicate that pest and beneficial arthropods are not reduced to the sanie dcgrec. Of the IS pest categories. 7 were little reduced by the ncctariless cultivar, whereas the remaining eight had reductions of 24 to 4% averaged for the 4-year study period. All 15 beneficial species categories showed a significant (P < 0.05) reduction on the nectariless cultivar (38 to 69% four year averaged reduction). these being affected to it greater degree than the pest. A disturbing question that remains is whcthcr the utility of nectarilehs cultivars for reducing pest pressures is afforded at comniercial levels of planting. .4s plot size increased froin 0. I I to 4.05 ha. the reduction on the ncctariless cultivar dccrcased from ca. 38 to 32%. Thc rate at which the differences between the nectaried and ncctariless cultivars decreases implies that the advantage with the nectariless cultivar might disappear with conimercial scale planting.\. Further large-scale trials would however be required for confirmation of this trend.

The reductions in population densities, both the 1975-1 976 to 1977-1 978 seasons and on November pests and beneficials, in nectariless cotton cultivars have 1st for thc 1978-1979 season. The resulting stand den- been demonstrated by several workers (George et al. sities were approximately 14 plants per m-row (I-m 1977, Laster and Meredith 1974. Lukefahr et al. 1965, spacing between rows) for the first 3 years and I I plants Maxwell et al. 1976). per ni-row for the 1978-1979 season. The crop was The small plot sizes used in such experiments do. furrow irrigated. nitrogen was applied before planting however, lend the interpretation of results to some crit- at a rate of 120 kglha. and no insecticidcs or acaricidts icism. Undoubtedly, the relatively mobile species which were used during the 4-year period. Plot size for each inhabit cotton primarily settle, feed, and reproduce in cultivar was 0.11. 0.61, 2.43, and 4.05 ha with 3, 3, the not too distant fields of “normal” nectaried cotton I. and I replicate. respectively, for the 1975-1976 to adjacent to the nectariless plots. Such a situation is 1978-1979 seasons. equivalent to a large-scale “choice” experiment. with Both near-isogenic ‘Stoneville’ cultivars were further most of the individuals congregating and building up on imbedded in a 50-ha block of ‘Deltapine 16’ (nectaried) the more preferred host. The obvious implication is that cotton. Sampling in each year commenced 3 weeks after in the absence of alternative nectaried cotton, as would plant emergence and was carried out by (I) whole-plant occur if the nectariless cotton were grown commercially visual examination of 1 -m-row lengths, which gave den- on a large scale, the pests and beneficials might occupy sity estimates largely for lepidopterous egg and larval the nectariless cotton with no resultant damage reduction populations, and by (2) a D-Vac sampler equipped with over that normally encountered with nectaried cotton. It a 25-cm-diameter cone (Dietrich 1961) which gave es- is even conceivable that, were beneficial arthropods more timates of plant bugs and most beneficial popu- sensitive to host plant-related environmental changes, a lations. D-Vac samples were taken by vacuuming the reduction in their numbers might still occur with a large- entire surface of a plant while slowly moving down the scale planting and that, in the absence of, or an only row (I-m-row sample lengths). All samples were taken slight reduction in, pest numbers, the net effect would in the center of the plots, twice per week, with 20 sam- be an increase in damage. ples per cultivar on each sampling day. All D-Vac sam- This paper reports the results of a 4-year study which ples were fumigated with ethyl acetate, and were examined the effect of these and other factors on the identified and counted in the laboratory. relative attractiveness of near-isogenic nectaried and The following factors were examined in comparing nectariless cotton to 30 categories of pest and beneficial the abundance of insects on the nectaried and nectariless arthropods. cultivars. (I) Arthropod feeding characterization: (a) pests. (b) bcneficials (predators and parasites). (2) Arthropod Materials and Methods density. (3) Crop phenological stage. (4) Plot size. The experimental site was at Forest Hill in the Lock- Regression analyses were used to compare the relative yer Valley, Queensland, Australia. Thc cultivars used attractiveness of the two cultivars for each category, were the nectaried and nectariless near-isogenic lines of where the number (weekly average) of each category on ‘Stoneville 7A.’ Seeds were planted on October 30th for the nectariless (Y) was rcgressed on those on the nec- taried (X) for each sampling date. The regressions were

lRcccivcd lor publication 5 April 19x2. Thi+ rc\carch uii\ conJuctcJ while forced through the origin (0.0) providing a regression I.K.? -M. was a sludcnt at the Univcr+ity olQuccn\lnnd. Athiralia. coefficient (b) which is an estimate of the relative at- -Piesent address: Dcpt. of Enioniology. Oil Paliii Rsmrch Centre. P.O. BOY 74. Kndc. Ghana. tractiveness of the two cultivars which can also be ap- ’hqt. of Entoinology. Univcrsity oi Cahiornia. IAIVI\. C‘A. MblO ~~ea\c send reprint rcquc\t\ to this iiddrc\s proximated by the ratio of thc numbers found on both

349 350 ENVIRONOMENTALENTOMOLOGY Vol. 12, no. 2 cultivars. The validity of using the forced regression 4-year study period. The values for each category rep- (Steel and Torrie 1960, Zar 1974) was determined for resent a summation of the average number, for all 4 all categories studied. years, of individual insects recorded during the 20 weeks of sampling. The Heliothis spp. were the most abundant Results and Discussion of the insects recorded, which is relevant considering Table I provides a list of the 30 categories of arthro- their importance as pests of cotton. For the 30 cate- pods and their cumulative numbers recorded during the gories, 23 of the regressions showed a significantly (P

Table I.-Cumulative average numbers (per m) recorded for the 30 arthropod categories (25 species) totaled for the 1975-1976 through 1978-1979 seasons for 'Stoneville 7A' nectaried (N+) and nectariless (N-) cotton, Lockyer Valley, Queensland, Australia

Age Sampling Pest % Reduction Species class" method status!' ZN + ZN - on the N-<

Lepidoptera Noctuidae Heliothis spp. E v P I .046.0 655.4 37 L V P 493.4 288.6 42 Eurius huegeli (Rogenhofer) L V P 58.0 44.2 24 Anomisfluvu F. L V P 139.8 137.8 I NS Tonricidae Crocidosemu plehicinu Zeller E V P 943.4 682.6 28 L V P 361.2 243.8 33 Pyralidae Dichocrocis punctiferulis (GueneC) L V P 39.4 37.4 5 NS Anthocoridb A D B 209.6 99.8 52 Miridae Derueocoris sipatus (Distant) A D B 360.0 178.0 51 Cumpylommu lividu Reuter N,A D P' 363.2 230.4 37 Megucoelum modestum (StHI) N.A D P' 278.6 200.2 28 Lygaeidae Geocoris lubru (Kirkdldy) N.A D B 59.6 35.0 41 Okycurenits spp. A V P 682.0 543.0 20 NS Nabidae Nubis cupsiformis Germar D B 209.6 98.8 53 Dysdercus sidue Montrouzier A V P 215.6 216.0 -0 NS Scutelleridae Tectocoris diopthalmits (Thunberg) N.A V P 597.8 569.8 5 NS Pentatomidae Nezuru viridulu (L.) N.A D P 105.0 60.6 42 Neuroptera Chrysopidae Chrysopa spp. E V B 225.4 69.8 69 L D B 148.0 51.6 65 A V B 132.6 52.6 61 Coleoptera Melyridae Luius bellalus GuCrin A D B 204.8 97.0 53 Coccinellidae Coccinellu repundu Thunberg A D B 237.6 141.2 41 Veruniu frenula Erichson A D B 300.6 166.0 45 Hurmoniu urcuutu (F.) A D I 190.0 172.6 9 NS Coelophoru inueyuulis (F.) A D I 145.4 130.8 IO NS Hymenoptera lchneumonids A D B 212.8 110.4 48 Brnconidae Microplitis spp. A D B 173.4 83.2 52 Araneida Oxyopidae Oxyopes spp. LA D B 256.0 151.8 41 Clubionidae Chirucunthium diversum Koch L.A D B 444.8 264.4 41 Theridiide Achueuruneu veriiculrtu (Urquhan) LA D B 245.8 152.8 38

"E. Egg; L; larva; N. nymph: A, adult. "B. Beneficials; I, innocuous pollen feeders (can function as faculrative predators): P. pests 'Significantly fewer (P < 0.05) on nectariless unless followed by NS (not significant). dFaculrative predators. April 1983 ADJEI-MAAFO AND WILSON: NECTARIED AND NECTARILESSCOTTON 35 1 < 0.05) greater number on the nectaried cultivar (Table I). A greater number was recorded in the nectariless cultivar for only one of the 30 species categories, this 5 PESTS being Dysdercus sidae (not significant, P > 0.05), a 4 late-season pest which feeds mostly on the developing seeds of maturing cotton bolls. Further breakdown of 3 1 the data showed only 3 (2.6%) of the 115 category x year comparisons (not all 30 categories were present wm2 every year) and only 256 (15.3%) of the 1673 category 0 WI x week x year comparisons with greater numbers on a the nectariless cultivar. Eighty-six of the I15 category cn x year regression comparisons had significantly (P < 0.05) fewer numbers on the nectariless cultivar, further ' 7 1 BENEFICIALS confirming its lower attractiveness (slope deviating sig- nificantly from 1 .0). Significant differences were not observed for the remaining 29 comparisons.

Pest and Beneficid Arthropods Of importance from both an ecological and pest man- agement context is the relative level of population suppression for pest and beneficial species. If beneficial species are suppressed relatively more than the associ- ated pests, it is theoretically conceivable that a greater amount of crop damage could result due to higher pest survival. Examining the percent reduction on the nec- tariless cultivar for each arthropod category for each year separately, 26 of the 27 lowest reductions were observed for seven of the categories illustrating the con- sistency in the relative attractiveness of the two culti- vars. Five of the seven categories which were least affected '10 REDUCTION ON THE NECTARILESS COTTON by the cultivar are pest species, whereas the remaining CULTIVAR (AVERAGE OF 4 YEARS DATA) two feed primarily on pollens. FIG. I .-Percent reduction on the 'Stoneville 7A' cultivar Of the 30 arthropod categories represented in the anal- for the pest and beneficial arthropods (averaged for 1975-1976 yses, 15 were beneficials and 15 were pest species or through 1978-1979 seasons). innocuous plant feeders (primarily pollen feeding adult coccinellids). Figure 1 shows a frequency histogram of the percent reduction of the nectariless for the 30 cate- gories with a distinction made between the pests and beneficials. The beneficials are obviously reduced to a 38 - greater extent than are the pests or innocuous plant feed- b ers (also see Table I). Fortunately, the Heliothis spp., the most important pests of cotton in Australia, were b 36 - reduced to a degree comparable to that observed for the beneficials. This was possibly due to the nectar-feeding habit of the adult moths. - 34 . Arthropod Density-Plant Stuge Of interest is whether the density of the pests and the stage of crop growth affect the relative attractiveness of the two cultivars. Unless nectariless cotton is less at- HECTARES OF NECTAR1LESS COTTON /PLOT tractive than nectaried cotton when exposed to low as well as high arthropod densities, little benefit can be FIG. 2.-Average percent reduction of arthropods on the expected, since control decision thresholds for most pests 'Stoneville 7A' ncctariless cultivar, as affected by plot size. are relatively low. Attractiveness as affected by stage of crop growth is important because cotton's susceptibility changes in a dynamic fashion as the crop matures (Blood high and low densities for each category for each week and Wilson 1978, Evenson 1969, Kincade et al. 1970, of sampling for the four seasons of the study. Low- and Wilson and Bishop, in press). To examine the effect of high-density and early- and late-season classifications crop stage and arthropod density, the data were broken were relative and not the same for all 30 categories, down into early and late season, and thcn further into since densities vary between categories, and some cat- 352 ENVIRONOMENTAI.ENIOMOLOGY Vol. 12, no. 2 cgorics arc present only rclativcly early or late in the trcnd was not apparent, indicating that plot size may not season. differentially affect the beneficial and pest arthropods. Analysis of the percent rcduction on the nectariless This still docs not answcr whether the pest reductions cultivar shows a highly significant category effect as observed for small plots of nectariless cotton will also cxpcctcd from Table I, as well as a highly significant carry over to large-scale plantings. Figure 2. in fact, (P < 0.005) drnsity cffcct. As density increases, the implies that this might- not be the case. and that further..~. ... ncctarilcss cultivar becomcs slightly less attractive (32.1% resenrch is required to determine whether nectariless reduction at relatively low densities and 37.2% reduc- cultivars rcccive fewer pests and pest-related damage tion at high densities). The effect of crop growth was with large-scalc commercial plantings. not significant (P 0.05). and docs not appcar to dif- > Acknowledgment ferentially affect the pests and bcneficials. Four of the 15 beneficial categories had a greatcr reduction in the We thank N. J. Thomson for the ‘Stoncville 7A’ nec- nectxilcss cultivar at low vs. high densities during early taried and nectariless seeds, A. Brimblecombe for his scason, and 5 of IS late season. ‘Thc smc pattern was gcncrous use of land. and thc University of Quecnsland also observed for the pests. for making this study possiblc. REFERENCES CITED Plot Size Blood, P. R. IS., and L. T. Wilson. 1978. Field validation of a crop’pcst manapenient descriptive niodcl. Proc. SIM- A continuing concem with plant resistancc studies is SIG Chif.. Canberra. Australia. pp. 9 1-94. whether the experimental design, in particular plot size. Dietrich, E. J. 1961. An improved backpack motor fan for allows a meaningful1 estimate of cultivar performance. suction sampling of insect populations. J. Econ. Entomol. Unfortunately, the limited quantity of experimental cul- 54: 394-395. tivar material and the hesitance of farmers to allow “un- Evenson, J. P. 1969. Effects of floral and terminal bud re- tested” cultivars to be planted on a large scale normally moval on the yield and structure of the cotton plant in the precludes the use of large-scale field testing. During thc Orcl Valley. North Western Australia. Cotton Grow. Rev. 4 years of this study, the plot size was increased from 46: 37-44. George, I). W., R. S. Slay, and D. L. Coudriet. 1977. Nec- 0.11 to 4.05 ha for the near-isogenic nectariless cultivar tarilcss and pubescent characteristics in cotton: effect on (each year embedded within a 50-ha block of commer- thc cabbage looper. J. Econ. Entomol. 70: 267-269. cial nectaried cotton). This represents a nearly 40-fold Kincade, R. T., M. L. Laster, and J. R. Brazzel. 1970. increase in plot size from the first to the fourth year of Effect of cotton yield of various levels of simulated damage the study and, although the plots were relatively small to squares and bolls. Ibid. 63: 613-615. compared with commercial fields, a definite decrease in Laster, M. L., and W. R. Meredith. 1974. Influence of the average percent reduction for the 30 arthropod cat- nectariless cotton on insect pest populations. MAFES Res. egories was apparent as plot size increased (Fig. 2). An Highlights. No. 1241. implication is that at commercial-scale plantings the rc- Lukefahr, M. J., D. F. Martin, and J. R. Meyer. 1965. Plant resistance to five lepidoptera attacking cotton. J. Econ. ductions observed for the experimental plots of thc nec- Entomol. 58: 516-518. tariless cotton may be lost. Testing of such a hypothesis Maxwell, F. G., M. F. Schuster, W. R. Meredith, and M. is difficult when plot size remains small, due to the L. Laster. 1976. Influence of the nectariless characteristics compounding interaction of (I) reductions in pest num- in cotton on harmful and beneficial insects. Symp. Biol. bers attributed to the relative attractiveness of the cul- Hung. 16: 157-161. tivar, and (2) the reduction due to beneficial arthropod Steel, R. G. D., and J. H. Torrie. 1960. Principles and pro- induced mortality. If the ratio of pests to beneficial is cedures of statistics. McGraw-Hill Book Co., New York. not affected by plot size, then relative beneficial induced 481 pp. Wilson, L. T., and A. L. Bishop. Responses of Deltapine 16 mortality is not likely to be affected by plot size, which cotton Gossyiitm hirsirritrn L. to simulated attacks by known in turn implies that damage reduction is proportional to populations of Heliothis larvae (Lepidoptera: Noctuidae) in the reduction in pest numbers. The ratio of (I - pro- a field experiment in Queensland, Australia. Prot. Ecol. (in portion reduction in pests)/( I - proportion reduction press) bencficials) for the 4-year period was I .S3, 1.54, 1.49. Zar, ,I. H. 1974. Biostatistical analysis. Prentice Hall, Engle- 1.58 (1975-1976 to 1978-1979, respectively). A clear wood Cliff>, N.J. 670 pp.