Adult Sorghum Midge (Diptera: Cecidomyiidae)Nonpreference for a Resistant Hybrid Sorghum

JOSÉ M. WAQUIL; GEORGE L. TEETES,ANDGARY C. PETERSON2

Department of Entomology, Texas A&M University,College Station, Texas 77843

WAQ, J.M.1986

J. Econ.Entorno!.79: 455-458 (1986)ABSTRACT Adult density and oviposition of sorghum midge, Contarinia sorghicola (Co-quíllett), on a resistant and susceptible sorghum, Sorghum bicolor (L.) Moench hybrid, wereassessedunder field conditions at College Station, Tex., during 1983 and 1984. Approximately33% more female midges visited flowering panicles of a susceptible hybrid than those of aresistant hybrid. Moreover, ca. 25%more female midges were collected on flowering paniclesof either hybríd that had not previously been infested with mídges compared with paniclesthat had already been exposed to midges. Number of eggs laid in spikelets by 20 cagedfemales per panicle, or by naturally occurring females in the field, was from 4- to 5-foldgreater for midges infesting a susceptible than a resistant hybríd. Differences in number ofeggs and egg-infested spikelets of the two hybrids fre consistent in a free-choice or no-choice situation. These results indicated that nonpre tence for oviposition was much moreimportant than nonpreference for adult visitation a a component of nonpreference resis-tance of the agronomically improved sorghum hybrid ATx2755 x Tx2767.

INSECTPESTSof sorghum reduce yield and in-crease production costs. The sorghum midge, Con-tarinia sorghicola (Coquillett), is one of the mostdamaging arthropod pests of sorghum, Sorghumbicolor (L.) Moench, worldwide (Young and Teetes1977). Sorghum grown between latitudes 40"N and40"S is always vulnerable to the sorghum midge.

Sorghum midge larvae are protected inside thesorghum spikelet during their development. Adultslive for 1 day and continuous daily infestation oc-curs in flowering sorghum fields. Consequently, in-secticidal control becomes expensive, and is oftenlargely ineffectíve when mídge density is high. Inaddition, sorghum is sensitive to some insecticidesand phytotoxicity often results (Meisch et al. 1970).

Cultural methods of control, such as destructionof wild hosts and host evasion by early, uniform,regional planting are the most effective controlstrategies (Hoelscher and Teetes 1981). The use ofresistant cultivars provides a promising midgemanagement tactic. Many lines of sorghum havebeen reported as resistant to sorghum midge. Bow-den (1965) and Harris (1970) reported that the'Nunaba' group of varieties was resistant becausethe glumes were long and díd not open duringanthesis. The same phenomenon was reported for'IS2663' and 'IS2660' sorghum lines by Berquist etaI. (1974).

Studying the response of eight lines of sorghumto midge infestation, Overman (1975) observed

1 Currentaddress:CNPMS-EMBRAPAc.P. 151, 35700 SeteLagoas,MG,Brazil.

•Currentaddress:TexasAgric.Exp.Stn.,Rt.3, Lubbock,TX77401.

051

that, although flowering panicles of 'AF-28' weremost frequently visited by midge females, theywere the least infested with eggs and consequentlyleast damaged. On the other hand, flowering pan-icles of the line 'SGIRL-MR-1' were visited lessoften than panicles of 'AF-28' but were moreheavily infested with eggs and consequently moredamaged. Based on the work of Overman (1975),it is possible to identify two components of adultnonpreference involved in the sorghum and sor-ghum midge interaction: adult nonpreference forvisitation and adult nonpreference for oviposition.Rossetto (1977) evaluated ca. 20 midge-resistantsorghum lines and concluded that nonpreferencefor oviposition was present to some degree in mostresistant lines. Wuensche (1980) reported thatcomparison of adult emergence from six sorghummidge-resistant lines failed to show differences inlevels of resistance. However, adults showed non-preference for visitation for the following linesin increasing order: 'SGIRL-MR-1', 'AF-28','TAM428', 'SC423', and 'TAM2566'. In addition,a significant reduction was found in the numberof larvae and pupae infesting spikelets of'TAM2566' and 'SGIRL-MR-1'. Evaluating yields,adult emergence, and caryopsis development,Wuensche (1980) found that tolerance was not animportant component of sorghum resistance tosorghum midge.

Melton and Teetes (1984) reported lower midgedensities on flowering panicles of two resistant hy-brids (both derived from 'T AM2566') than on pan-icles of a susceptible hybrid. Also, they reported a40-60% reduction in number of progeny thatemerged from spikelets of the resistant hybríds

455

456 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 79, no. 2

Table 1. Comparison of lhe mean number of sorghum midge females lrapped per panicle of a resístant andsusceptihle sorghum hybrid under Iree-choiee situation on seven occasions

X no. of fernales per panicle

Hybrid 1983a 1984b

18 [uly 20 July 31 [uly 6 Aug. 7 Aug. 3 Aug. 9 Aug. ..Resistant 17.1a 14.5a 7.2a 47.3a 74.3a 18.7a 106.2aSusceptible 25.1b 41.1b 16.4b 65.7b 146.1b 30.4b 148.7b

a Least-square rneans within colurnns followed by the sarne letter are not sígnífícantly dífferent (P > 0.05; Student's t test).b Means within colurnns followed by the same letter are not significantly dífferent (P > 0.05; Duncan's [1955] rnultiple range test).

compared to the susceptible check, regardless ofmidge density. For the common sorghum hybridsgrown in Texas, the economic threshold level forsorghum midge is an average of one ovipositingfemale per flowering panicle (Bottrell1971); how-ever, for improved midge-resistant hybrids theeconomic threshold level has been shown to be5-fold greater (Hallman 1982).

The objective of our research was to comparenonpreference for visitation and oviposition onflowering panicles of an agronomically improvedmidge-resistant and a midge-susceptible sorghumhybrid,

MateriaIs and Methods

Field experiments were conducted in BurlesonCounty, Tex., at the Texas Agricultural Experi-ment Station Research Farm near College Station.The hybrids used were: the midge-susceptiblecheck ATx2752 x Tx430 and the midge-resistantATx2755 x Tx2767 (Johnson et a!. 1982). Both hy-brids were planted on three occasions in four-rowplots (1 m wide, 6.1 m long) and replicated fourtimes. The cultural practices used were those rec-ommended for the region. To ensure that midgeinfestation occurred at the desired time, emergingpanicles at the boot stage of plant developmentwere individually covered with fine-mesh plasticcages sealed at the base with stripes of velcro sewedon the inside (Baxendale 1983). Cages were re-moved at the 50% panicle flowering stage to allowinfestation.

Adult Nonpreference for Visitation. The num-ber of sorghum midge adult females visiting pan-icles of the resistant and susceptible sorghum hy-brids was recorded on several occasions in 1983and 1984. In addition, the number of adult sor-ghum midges visiting uninfested and previouslymidge-infested flowering panicles of both hybridswas determined in 1984. Ten flowering paniclesof each hybríd, exposed to midge infestation, wererandomly chosen between plots in 1983 and withinplots in 1984. Each panicle was covered with aplastic bag (35 by 25 em), then excised and trans-ported to the laboratory where the number ofadults was counted. This experiment was con-ducted on five occasions in 1983 and two occasionsin 1984.

Adult Nonpreference for Oviposition. A com-parison of the number of eggs laid by sorghummidges in spikelets of the resistant and susceptiblesorghum hybríds and the number of egg-infestedspikelets was made in 1983 and 1984. Natural andartificial midge infestations were used. Natural in-festation of panicles was accomplished by remov-ing cages from panicles at the 50% flower stage,allowing panicle exposure to naturally occurringmidges from 0800 to 1600 hours (CDT) in 1983.Since midge density could not be manipulated un-der natural infestation, the interval of time thatflowering panicles were exposed to infestation wasmanipulated to expose the resistant hybrid to midgeoviposition pressure higher than that of the sus-ceptible hybrid. Panicles of the resistant hybridwere exposed to natural midge infestation for 3 hwhile panicles of the susceptible hybrid were ex-posed for only 1 h. Panicles were artifícially in-fested using the same kind of cages used for pro-tection. Twenty adult female midges per paniclewere introduced into cages from 0800 to 1000 hours(CDT) (Melton 1982).

One day following infestation, random samplesof one rachis branch from a section of the paniclethat was flowering the previous day were collect-ed, brought to the laboratory, and stored in afreezer for later spikelet dissection. Spikelets werechosen at random from the collected rachisbranches, then dissected under a stereomicroscopeto count eggs and number of infested spikelets. In1983, on six occasions, rachis branches were col-lected from five plants of each hybrid and bulked.A sample of 50 spikelets was dissected. In 1984,on three occasions, rachis branches were collectedfrom eight panicles of each hybrid and 25 spikeletsper panicle were dissected. To obtain a better ap-proximation of a normal distribution, logarithmsof the number of eggs and square root of the num-ber of infested spikelets were used for the statis-tical analysis. Significant dífferences (P < 0.05)were determined using analysis of variance, Stu-dent's t test, and Duncans (1955) multiple rangetesto

Results and Discussion

Adult Nonpreference for Visitation. The meannumber of midge females per panicle of the sus-

April1986 W AQUIL ETAL.: SORGHUM MmGE NONPREFERENCES 457

Table 2. Comparison of adult midge density on flow-ering panicles of a resistant and susceptible sorghum hy-brid in three independent trials

Treatmentf no. of adult midges per panicle

Susceptible" Besístant- Both hybríds''

UninfestedInfested

11.4a 19.1a 46.6a8.0b 14.2b 107.8b

a Least-square means within columns followed by the same let-ter are not significantly different (P > 0.05; Student's t test).

b Means within columns followed by the same letter are notsignificantly different (P > 0.05; Duncan's [1955] multiple rangetest).

ceptible hybríd was significantly (P < 0.05) higherthan the number per panicle of the resistant hy-brid (Table 1). Adult midge densities varied asmuch as lO-fold during the density assessment pe-riod. In general, l.4- to 2.9-fold more females werepresent on flowering panicles of the susceptiblehybrid than on panicles of the resistant hybrid,The results of these 2-year experiments indicatedsome degree of adult midge nonpreference for vis-itation to the flowering panicles of the resistanthybríd when susceptible sorghum was available.These results agree with severaI earlier reports(Wiseman and McMillian 1970).

The number of visiting adult females on flow-ering panicles of the resistant or susceptible hybridpreviously exposed to midge adults was signífí-cantly (P < 0.05) lower than the number on flow-eringpanicles that had not been previously midgeexposed or infested (Table 2). No significant (P >0.05) interaction effect was detected betweenmidge infestation and hybríd. Midge density wassignificantly (P < 0.05) higher on previously un-infested (unexposed) panicles than on panicles pre-viously infested by midges, regardless of hybrid,

Adult Nonpreference for Oviposition. Follow-ing natural infestation, the mean number of sor-ghum midge eggs was significantly (P < 0.05)greater in spikelets of the susceptible hybrid thanin spikelets of the resistant hybríd (Table 3). Thenumber of infested spikelets of the susceptible hy-

Table 3. Comparison of nonpreference for ovipositionof sorghum midges under natural infestation of resistantand susceptible sorghum, 1983 and 1984

x no. per sample

Hybríd E~g Infested spikelet

1983a 1983a

SusceptibleResistant

134.7a29.0b

63.6a26.7b

32.9a15.1b

18.3a11.8b

Statistical test performed with the logarithm (no. eggs) andsquare root plus 0.5 (no. infested spikelets) of original value. Means(1983) and least-square means (1984) within columns followedby the same letter are not significantly dífferent (P > 0.05; Dun-can's [1955] multiple range test and Student's t test, respectively).

a Flowering panicles of both hybríds were exposed to midgesfor a full day.

b Flowering panicles of the susceptible hybríd were exposed tomidges for 1 h and the resistant hybrid for 3 h.

Table 4. Comparison of nonpreference for ovipositionof sorghum midges under artificial infestation of a resis-tant and susceptible sorghum

X no. per sample

Hybrid Infested spikeletEgg

1983 1984 1983 1984

SusceptibleResistant

29.2a4.9b

21.0a5.1b

15.8a3.6b

11.6a3.7b

Means (1983) and least-square means (1984) within columnsfollowed by the same letter are not significantly different (P >0.05; Duncan's [1955] multiple range test and Student's t test,respectively).

brid was also significantly (P < 0.05) greater thanthose of the resistant hybrid in both years. Thedifference in the number of eggs laid in the spíke-lets, as well as the number of infested spikelets ofthe resistant and susceptible hybrids, was not asgreat in 1984 (2-fold) as it was in 1983 (5-fold) dueto infestation time differences.

The smaller number of eggs found in spikeletsof the resistant hybríd following natural infesta-tion could be a result of the lower visiting midgedensity on the panicles of the resistant hybríd,However, the difference in number of adults vis-iting panicles of the two hybríds was not as greatas the difference in egg infestation levels. Theseresults indicated the existence of another compo-nent of nonpreference resistance besides adultnonpreference for visitation. Following artificialinfestation, the mean number of sorghum midgeeggs laid between the glumes as well as the num-ber of infested spikelets were significantly (P <0.05) greater for the susceptible hybrid than theresistant hybrid (Table 4). About 5-fold more eggsand 4-fold more infested spikelets were recordedfor the susceptible hybrid than the resistant hy-brid.

The significantly larger egg infestation of thesusceptible hybrid compared with the resistant hy-brid, under the same adult density, indicated thatnonpreference for oviposition is a component ofresistance in addition to adult nonpreference forvisitation. Although this nonpreference compo-nent is referred to as nonpreference (Painter 1951),the adult midge females may not have been ableto lay as many eggs in the spikelets of the resistanthybrid as in the spikelets of the susceptible hybrídor may have been deterred from laying. When theresults of experiments involving adult nonprefer-ence for visitation were compared with that foroviposition under natural and artificial infestation,some important implications were apparent. Midgeadult nonpreference for oviposition, ar inability tooviposit, would be expected to be additive to non-preference for visitation. However, this additiveeffect did not occur. The dífference in egg infes-tation leveI between the resistant and susceptiblehybrid was the same (ca. 5-fold) following nat-ural or artificial infestation. Natural infestationinvolves both resistance components (adult non-

458 JOURNAL OF ECONOMIC ENTOMOLOGY VoI. 79, no. 2

preference for visitation and oviposition) whileartificial infestation involves only nonpreferencefor oviposition. At least two hypotheses can be pro-posed. Since adult midge females can distinguishinfested panicles from uninfested panicles, as dis-cussed previously, they may switch from paniclesof the susceptible hybrid, after heavy oviposition,to the resistant hybrid, reducing the difference inegg infestation level between the two hybrids un-der natural infestation. Secondly, adult midge ovi-position decreases when adults are caged on pan-icles (Hallman 1982). Caged females comparedwith uncaged females lay fewer eggs in the spike-lets of resistant sorghum than in the susceptiblesorghum. If it is more difficult for females to layeggs in flowers of the resistant hybrid than in thesusceptible hybríd, the disrupting effect caused bythe cage could further reduce oviposition in spike-lets of the resistant compared with the susceptiblehybrid.

The results of these experiments indicated theexistence of some degree of adult midge nonpref-erence for visitation to flowering panicles of theresistant hybrid. Also, adult midge density waslower on previously infested flowering panicles thanuninfested panicles, independent of hybrid. Four-to 5-fold more eggs were laid in spikelets of thesusceptible hybrid than in the resistant hybríd un-der a no-choice situation. Comparing the resultsof number of eggs laid in spikelets of both hybridsunder a free-choice situation with the number laidin a no-choice situation, we conclude that nonpref-erence for oviposition is the major mechanism ofresistance, reducing the number of eggs laid andegg-infested spikelets of the agronomically im-proved midge-resistant hybríd ATx2755 x Tx2767.

Acknowledgment

This research is approved as TA20918 by Director,Tex. Agric. Exp. Stn. and was supported in part byEMBRAPA and by grant AID/DSAN/XII/G-0149 frornthe Agency for International Developrnent, ProjectH-6216. We thank Nicole Housser for her invaluableassistance in collecting data.

References Cited

Baxendale, F. P. 1983. Modeling seasonal develop-ment of the sorghum midge, Contarinia sorghicola(Coquillett). Ph.D. dissertation, Texas A&M Univer-sity, College Station.

Berquist, R. R., P. Rotar, and W. C. Mitchell. 1974.Midge and anthracnose head blight resistance in sor-ghum. Trop. Agric. 51: 431-535.

Bottrell, D. G. 1971. Entomological advances on sor-ghum productíon, pp. 28-40. In Grain sorghum re-search in Texas ... 1970. Tex. Agric. Exp. Stn. PR-2938-2949.

Bowden, J. 1965. Sorghum midge, Contarinia sorgh-icola (Coq.), and other causes of grain sorghum lossin Ghana. Bul!. Entorno!. Res. 56: 169-189.

Duncan, D. B. 1955. Multiple range and multiple Ftests. Biometrics 11: 1-41.

Hallman, G. J. 1982. Relationship of sorghum mídge,Contarinia sorghicola (Coquíllett), density and dam-age to resistant and susceptible sorghum hybríds,Ph.D. díssertatíon, Texas A&M Uníversíty, CollegeStation.

Harris, K. M. 1970. The sorghum midge. PANS (PestArtic. News Summ.) .16: 36-42.

Hoelscher, C. E., and G. L. Teetes. 1981. Insect andmites pests of sorghum-management approaches.Tex. Agric. Ext. Ser. Pub!. B-1220.

Johnson, J. W., G. L. Teetes, D. T. Rosenow, B. R.Wiseman, and J. M. Phillips. 1982. Registrationof 28 midge resistant sorghum germplasm lines. CropSci. 22: 1273.

Meisch, M. V., G. L. Teetes, N. M. Randolph, and A.J. Bockholt. 1970. Phytotoxic effects of insecti-cides on six varieties of grain sorghum. J. Econ. Ento-rno!. 63: 1516-1517.

Melton, K. D. 1982. Effects of resistant grain sor-ghum hybríds on the biology of the sorghum rnidge,Contarinia sorghicola (Coquillett). M.S. thesis, TexasA&M University, College Station.

Melton, K. D., and G. L. Teetes. 1984. Effects ofresistant sorghum hybrid on sorghum mídge (Dip-tera: Cecidornyiidae) biology. J. Econ. Entorno!. 77:629-631.

Overman, J. L. 1975. Some perspecti ves on insectproblems of sorghum in Brazil, pp. 322-329. In Proc.International Sorghum Workshop, University ofPuerto Rico, Mayaguez.

Painter, R. H. 1951. Insect resistance in crop plants.Macmillan, New York.

Rossetto, C. J. 1977. Tipos de resistencia de sorgo,Sorghum bicolor L. Moench, a Contarinia sorghi-cola (Coquillett, 1898). Tese de Livre Docente, Uni-versidade Estadual Paulista "Julio de Medquita Fil-ho," Jaboticabal Brazi!.

Wiseman, B. R., and W. W. McMillian. 1970. Pref-erence of sorghurn rnidge among selected sorghumlines, with notes on overwintering midges and par-asite emergence. U.S. Dep. Agric. Prod. Res. Rep.122.

Wuensche, A. L. 1980. An assessment of plant resis-tance to the sorghum midge, Contarinia sorghicola,in selected lines of Sorghum bicolor. Ph.D. disserta-tion, Texas A&M University, College Station.

Young, W. R., and G. L. Teetes. 1977. Sorghumentomology. Annu. Rev. Entorno!. 22: 193-218.

\

"

Received for publication 30 August 1985; accepted25 November 1985.