Document generated on 09/27/2021 7:41 p.m.

Phytoprotection

Evaluation of resistance to the , Acyrtosiphon pisum [Homoptera : ] – Methodological aspects to improve a standardized speedling test Évaluation de la résistance de la luzerne au puceron du pois, Acyrtosiphon pisum [Homoptera : Aphididae] - Améliorations méthodologiques d’un test « plantule » standardisé C. Girousse, R. Bournoville and I. Badenhausser

Volume 79, Number 3, 1998 Article abstract This study proposes a guide for the design of experiments to test alfalfa URI: https://id.erudit.org/iderudit/706142ar (Medicago sativa) for resistance to pea aphid infestation ( pisum). DOI: https://doi.org/10.7202/706142ar This test was conducted in controlled conditions on alfalfa seedlings. For the infestation, aphid population maintained on alfalfa was found to be more See table of contents efficient than an aphid population reared on broad bean. When comparing alfalfa , a non-choice test gave the same results as a choice test, that was more difficult to perform. When infesting a unit of 54 seedlings at the st th Publisher(s) cotyledon stage on the 1 and 5 day of the experiment, 360 mg compared with 180 mg and 540 mg , led to the best compromise between levels of Société de protection des plantes du Québec (SPPQ)l infestation and aphid stock culture availability. Infestation was stopped when more than 60% of susceptible seedlings were wilted or dead. Under ISSN these conditions, we calculated the number of replicates necessary to obtain a fixed level of difference. Six units per cultivar would distinguish between 0031-9511 (print) cultivars differing from 20% in their seedling mortality. 1710-1603 (digital)

Explore this journal

Cite this article Girousse, C., Bournoville, R. & Badenhausser, I. (1998). Evaluation of alfalfa resistance to the pea aphid, Acyrtosiphon pisum [Homoptera : Aphididae] – Methodological aspects to improve a standardized speedling test. Phytoprotection, 79(3), 139–148. https://doi.org/10.7202/706142ar

La société de protection des plantes du Québec, 1998 This document is protected by copyright law. Use of the services of Érudit (including reproduction) is subject to its terms and conditions, which can be viewed online. https://apropos.erudit.org/en/users/policy-on-use/

This article is disseminated and preserved by Érudit. Érudit is a non-profit inter-university consortium of the Université de Montréal, Université Laval, and the Université du Québec à Montréal. Its mission is to promote and disseminate research. https://www.erudit.org/en/ Evaluation of a If a If a résistance to the pea aphid, Acyrthosiphon pisum [Homoptera: Aphididae] Methodological aspects to improve a standardized seedling test

Christine Girousse, René Bournoville and Isabelle Badenhausser1

Received 1998-10-15; accepted 1999-03-15 PHYTOPROTECTION 79 : 139-148

This study proposes a guide for the design of experiments to test alfalfa {Medicago sativa) for résistance to pea aphid infestation (Acyrthosiphon pisum). This test was conducted in controlled conditions on alfalfa seed- lings. For the infestation, aphid population maintained on alfalfa was found to be more efficient than an aphid population reared on broad bean. When comparing alfalfa cultivars, a non-choice test gave the same results as a choice test, that was more difficult to perform. When infesting a unit of 54 seedlings at the cotylédon stage on the 1st and 5th day of the experiment, 360 mg compared with 180 mg and 540 mg aphids, led to the best com­ promise between levels of infestation and aphid stock culture availability. Infestation was stopped when more than 60% of susceptible cultivar seed­ lings were wilted or dead. Under thèse conditions, we calculated the num- ber of replicates necessary to obtain a fixed level of différence. Six units per cultivar would distinguish between cultivars differing from 20% in their seedling mortality. [Évaluation de la résistance de la luzerne au puceron du pois, Acyrthosi­ phon pisum [Homoptera : Aphididae] - Améliorations méthodologiques d'un test «plantule» standardisé] Cette étude concerne des éléments de standardisation pour mettre en oeuvre un test d'évaluation de la résistance variétale de la luzerne (Me­ dicago sativa) au puceron du pois (Acyrthosiphon pisum). Réalisé en con­ ditions contrôlées, il évalue la résistance de plantules à l'infestation aphi- dienne. Pour réaliser cette dernière, il est nécessaire de produire les pucerons sur des plantes de luzerne plutôt que sur des plantes de féverole, leur action sur la luzerne étant alors moindre. La disposition d'un seul cultivar (con­ dition de «non choix») donne les mêmes résultats que la disposition dans la même unité de plusieurs variétés (conditions de choix) et est plus facile à gérer. L'unité élémentaire, constituée de 54 plantules, est infestée avec 360 mg de pucerons lorsque les cotylédons s'ouvrent, ce que l'on renou­ velle après 5 jours. Cette dose conduit à des résultats optimums par rapport à des infestations réalisées dans les mêmes conditions avec 180 ou 540 mg. L'infestation est arrêtée lorsque 60 % des plantules d'un cultivar sensible

1. Laboratoire de Zoologie, Institut national de la Recherche agronomique, 86600 Lusignan, France

139 dépérissent ou meurent. Dans ces conditions, nous avons calculé le nom­ bre de répétitions nécessaires. Six unités élémentaires permettent alors d'apprécier une différence de mortalité entre cultivars de 20 %.

INTRODUCTION grow) in a test of résistance of alfal­ fa seedlings to PA ? (Experiment 1). The évaluation of plants for ré­ 2 - Is it possible to use a non-choice sistance may be based on plant dam­ test instead of a more complicated age or on insect responses to plants. choice test when comparing alfalfa Although pea aphid (PA) Acyrthosiphon génotypes for résistance to PA ? (Ex­ pisum (Harris) [Homoptera : Aphididae] periment 2). is one of the main pests of alfalfa {Med- 3 - What are the minimum aphid bio­ icago sativa L.) in Europe (Bournoville masses and number of insect releas- 1976), there are few European studies es necessaryfordetecting résistance about méthodologies for evaluating of alfalfa seedlings ? (Experiment 3). alfalfa résistance to PA whatever the In fact, if infesting populations are underlying type of résistance, tolérance too high, moderate levels of résis­ or antibiosis (Bournoville 1980; Bourn­ tance may appear susceptible, oville and Comte 1977; Girousse and whereas when insect populations Bournoville 1994). In North America, are too low, the screening between various methods for the évaluation of résistant and susceptible cultivars alfalfa résistance to the PA hâve been may be unreliable (Smith et al. 1994). carried out, mainly under greenhouse conditions (Hackerott et al., 1963; Ku- Experiments were carried out on gler and Ratcliffe 1983; Nielson and alfalfa cultivars with well-known degree Lehman 1977;Ortman et al. 1960), some- of résistance in controlled conditions times in growth chambers (Kugler and as the expression of alfalfa résistance Ratcliffe 1983) or outside under screen to aphids could vary according to the cages (Carnahan et al. 1963; Smith and température as reported for the pea Peaden 1960). Studies about methods aphid (Isaak et al. 1963,1965), the spot- to test alfalfa résistance under field ted alfalfa aphid, Therioaphis maculata conditions hâve also been carried out [Homoptera : Aphididae] Buckton (Schalk in Australia because of the high level of et al. 1969) and the blue alfalfa aphid, aphid infestations in that country (Holt- [Homoptera : kamp and Clift 1993). The existence of Aphididae] Shinji (Summers 1988). several différent évaluation procédures for résistance is surprising, as breeding programs need efficient, simple and ac- MATERIAL AND METHODS curate screening methods (Smith et al. 1994) which should be standardized (Sorensen et al. 1988). Plants Alfalfa seeds were germinated on a wet The most récent standardization of a filter paper in Pétri dishes. After 3 to 4 method to test alfalfa résistance to the d, when the cotylédons were opening PA, at a seedling stage has been de- and young roots were about 1 cm long, scribed by Berberet et al. (1991). Nev- seedlings were planted in individual ertheless, some methodological élé­ peat pots (4 cm x 4 cm x 5 cm, Jiffy-pots ments remained to be demonstrated. Reg.) containing a soil mixture (sand, Our aim was then to clarify three of compost, végétal earth) which had been thèse éléments, running three différent sieved and heated to destroy pathogens experiments. and other seeds. A tray consisting of 54 1 - Is it relevant to use PA reared on (six rows x nine columns) peat-pots, faba bean L (instead of thereafter called an unit, was put in a alfalfa which is more difficult to screened cage (40 cm x 35 cm x 30 cm)

140 GIROUSSE ET AL. : RESISTANCE TEST OF ALFALFA TO THE PEA APHID

with a transparent opening. This allowed experiments, the infestations were seedling growth and watering and aphid stopped with a spray of deltamethrin population growth. Ail the experiments (37.5 mg L1) when more than 60% of were carried out in a growth chamber. 'Milfeuil' seedlings were dead or wilted The air température was maintained at and ail the observations were made 20°C ± 2°C with a photoperiod of 16:8 immediately after spraying. In ail ex­ (L:D) and a relative humidity varying periments, treatments were compared from 70% (L) to 90% (D). Light intensity using the proportions of dead plus was 95 jimol m2s1 at 75 cm from flu­ wilted seedlings per unit. The propor­ orescent tubes, i.e.,above the seedlings. tions of seedlings from the différent stages of development were used for Standard test for PA résistance (Ber- the interprétation of the results. beret et al. 1991 ) refered to 'CU F 101 ' as résistant and 'Moapa 69' or 'Vernal' as Expérimental design susceptible cultivars (US origin). In thèse Three experiments were carried out experiments, we used 'CUF 101' and a consecutively to define the standards French cultivar 'Milfeuil' as a suscepti­ of the test. The results of one experi- ble control. ment were considered when designing into the following one. An uninfested control was added in each experiment Aphids were collected in a alfalfa field to verify that the growth of the seed­ in the INRA expérimental station in lings was optimal. Lusignan (near Poitiers, France). This aphid population was reared either on Influence of the host plant for faba bean (cv. Aguadulce), or on the aphid production (Exp. 1) susceptible alfalfa cultivar 'Milfeuil' in a Two treatments were compared for each growth chamber at 21 °C, photoperiod cultivar ('Milfeuil' and 'CUF 101') each 16:8 (L:D) at 75% RH. using one unit of 54 seedlings : infes- tation was realized with our aphid strain Infestations and observations reared either on faba bean or on alfalfa Plants were infested 1 d after trans- for several mo. Each seedling was in­ planting into peat-pots with a mixed fested using two adults of A. pisum population of nymphs and adults of A. (Berberet et al. 1991). pisum by sprinkling aphids as evenly as possible over the seedlings. As aphid Choice or non choice test (Exp. 2) weight is well-correlated with aphid Two treatments were compared. fecundity and therefore is a good mark- er for potential population dynamics, (1) 'CUF 10V and 'Milfeuil' planted in 20 randomly chosen wingless aphids of alternate rows and infested (two PA were weighed using a 10~5 g mi­ units of 54 seedlings) ; crobalance, just before the infestations. (2) 'CUF 101' or 'Milfeuil' planted each It allowed us to check whether ail the in separated units and infested (one experiments were conducted with com­ unit per cultivar). parable presumed aphid performanc­ es. Two adults of A. pisum (reared on alfalfa) per seedling were used for the In testing seedling résistance to PA, first infestation. A second infestation our observations concerned only alfal­ was carried out 7 d later using one A. fa seedlings because our design did pisum adult per seedling. not allow for recording aphid perfor­ mance, such as aphid mortality or fe­ Aphid biomass (Exp. 3) cundity. Each seedling was rated ac- As aphids are easier to weigh than to cording to a 3-class System : dead, wilted isolate and to count, we tried to déter­ or healthy. The différent stages of de- mine an infesting biomass of aphids in velopment of the healthy seedlings were place of a number of adults as previ- described as their number of leaves : C : ously used. In terms of biomass, an cotylédons; L0 : the first unifoliate leaf; infestation with 180 mg per unit was Li : i trifoliate leaves. Throughout the équivalent to an infestation carried out

141 with two adult aphids per plant (as the Comparisons of the proportions mean weight of an adult is 1.7 mg). of a defined class Then, three treatments were compared Thecomparison oftwo proportions was each using two units planted with 'Mil- conducted using the exact test (Dagne- feuil' and two units planted with 'CUF lie, 1986) based on the probability term 101' in a non-choice design. The initial (Pr) for the hypergeometric distribution. biomass of aphids added to the treat­ Under the hypothesis of the equality of ments was as follows : 1) 180 mg aphids; proportions, Pr gives the probability to 2) 360 mg aphids; 3) 540 mg aphids encounter a more abnormal frequency reared on alfalfa. A second infestation distribution than observed. with the same biomass was realized 4 d later. RESULTS Statistical analysis

Estimation of the proportions (p) Exp. 1 - Influence of the host and their standard errors (SE(p)) plant for aphid reproduction In the case where there was one unit of Whatever the cultivar, the effect of the m seedlings per treatment, p and SE(p) aphids reared on alfalfa was significantly were calculated as follows : différent from that of the aphids reared on faba bean (Pr=0.014 for 'CUF 10V and Pr=1 1011for 'Milfeuil'). When in­ m festation was carried out with aphids and reared on faba bean, the proportion of , [91 SE(p) = yp(1-p)/m wilted plus dead seedlings was nil for 'CUF 101" and 0.02 for 'Milfeuil' (Table 1). When infestation was carried out where yisthe numberof seedlings that with aphids reared on alfalfa, this pro­ fall into a defined class (for example portion was higher and consistent with dead seedlings). that of previous tests where the level of In the case where there were n units susceptibility was about 30% of dead of m. seedllings per treatment, p and seedlings for 'CUF 101' (Berberet et al. SE(p) were calculated according to 1991), and 70% for 'Milfeuil'. Cochran (1977) : Although there was no seedling n mortality when aphids were reared on faba bean, there was a delay in phenol- J4- [3] ogy when compared with the uninfest- P= ed control where almost ail the seed­ lings developed one trifoliate leaf (Table / = i 1). This delay was more important when and seedlings were infested with aphids reared on alfalfa. The phenology re- vealed différences between 'CUF 101' 2 2 SE(p) = yiK; -2pIyfm,+ p Sm,'r41 and 'Milfeuil', with 'CUF 101' seedlings (n - 1)n(ImV/i)2 being more developed than those of 'Milfeuil'. The conclusion is that, what­ ever the alfalfa cultivar (susceptible or résistant), aphid impact on alfalfa seed­ § where yjsthe numberof seedlings that lings dépends on the host plant used p fall into a defined class within the mj for aphid reproduction; in the following UJ seedlings in unit i. (In most cases, ml experiments we used aphids reared on O was 54 but in few cases, seedlings hâve alfalfa. g been eliminated due to extra mortality).

142 GIROUSSE ETAL : RESISTANCE TEST OF ALFALFA TO THE PEA APHID

Table 1. Influence of the host-plant species for aphid reproduction (A = alfalfa ; B = faba bean) on aphid damaging effect. Proportions p and standard error SE(p) in parenthesis of healthy, wilted and dead seedlings

'CUF 101' infested with aphids 'Milfeuil' infested with aphids from from Seedling class A B A B Un infested control Healthy3 0.88 (0.05) 1.00 (0.00) 0.35 (0.07) 0.98 (0.02) 1.00 (0.00) C 0.10 (0.04) 0.06 (0.03) 0.13 ( 0.05) 0.14 (0.05) 0.02 (0.02) LO 0.72 (0.07) 0.25 (0.06) 0.22 (0.06) 0.27 (0.06) 0.06 (0.03) L1 0.06 (0.03) 0.69 (0.07) 0.00 (0.00) 0.57 (0.07) 0.92 (0.04) Wilted 0.12 (0.05) 0.00 (0.00) 0.03 (0.02) 0.02 (0.02) 0.00 (0.00) Dead 0.00 (0.00) 0.00 (0.00) 0.62 (0.07) 0.00 (0.00) 0.00 (0.00) Wilted + dead 0.12 (0.05) 0.00 (0.00) 0.65 (0.07) 0.02 (0.02) 0.00 (0.00) a For healthy seedlings, the proportions of the seedlings at différent developmental stages are detailed : C: cotylédons; L0: unifoliate leaves; L1: first trifoliate leaves.

Table 2. Comparison of choice / non-choice test. Proportions p and standard error SE(p) in parenthesis of healthy, wilted and dead seedlings

'CUF 101' 'Milfeuil' Seedling class Choice Non-choice Choice Non-choice Uninfested control Healthy3 0.84 (0.05) 0.80 (0.06) 0.17 (0.05) 0.19 (0.05) 0.92 (0.04) C 0.00 (0.00) 0.06 (0.03) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) L0 0.09 (0.04) 0.14 (0.05) 0.13 (0.05) 0.17 (0.05) 0.00 (0.00) L1 0.08 (0.04) 0.04 (0.03) 0.02 (0.02) 0.02 (0.02) 0.02 (0.02) L2 0.15 (0.05) 0.18 (0.05) 0.02 (0.02) 0.00 (0.00) 0.00 (0.00) L3 0.30 (0.06) 0.28 (0.06) 0.00 (0.00) 0.00 (0.00) 0.61 (0.07) L4 0.22 (0.06) 0.10 (0.04) 0.00 (0.00) 0.00 (0.00) 0.29 (0.06) Wilted 0.09 (0.04) 0.14 (0.05) 0.46 (0.07) 0.62 (0.07) 0.04 (0.03) Dead 0.07 (0.04) 0.06 (0.03) 0.37 (0.07) 0.19 (0.05) 0.04 (0.03) Wilted + dead 0.16 (0.05) 0.20 (0.06) 0.83 (0.05) 0.81 (0.05) 0.08 (0.04) a For healthy seedlings, the proportions of the seedlings at différent developmental stages are detailed : C: cotylédons; L0: unifoliate leaves; Li: ith trifoliate leaves.

Exp. 2 - Choice or non choice test of wilted plus dead seedlings was high- As there was no différence in the pro­ er because of longer duration of the portions of wilted and dead seedlings infestation : about 20% for 'CUF 101' between the two replicatesoftreatment and about 80% for 'Milfeuil' in a non- 1 (choice test) (Pr=0.52 for 'CUF 101' choice test. and Pr=0.195 for 'Milfeuil'), data were pooled. The proportion of wilted plus The developmental stages of surviv- dead seedlings are reported for both ing seedlings for both cultivars and cultivars and tests (Table 2). There was devices are reported in Table 2. Gener- no différence between the two kinds of ally, the seedlings were more devel- test whatever the cultivar (Pr=0.463 for oped in this experiment than in the 'CUF 101' and Pr=0.693 for 'Milfeuil'). previous one because of the longer The same différence as in experiment duration of the experiment (12 more d). 1 was observed in the susceptibility of The différence in susceptibility be­ the two cultivars, even if the proportion tween 'CUF 101 ' and 'Milfeuil' was also

143 expressed in the developmental stages DISCUSSION of surviving seediings in a non-choice test: morethan 56% of'CUF 101' seed­ iings developed more than one trifoli- Standardization of seedling test ate leaf, whereas no seedling of 'Mil- of alfalfa résistance to the pea feuil' had more than one trifoliate leaf. aphid Apart from this, 90% of the uninfested Exp. 1. In the résistance tests described control seediings had more than two in the literature, the broad bean is often trifoliate leaves. In conséquence, as a used because of its convenience to be non choice device is easier to set up planted and maintained in controlled than a choice device, a non choice test conditions in order to breed large was prefered. amounts of PA. For example, Ellsbury and Nielson (1981) reported that faba Exp. 3 - Aphid biomass bean is a more favourable host plant As there was no différence for each for the pea aphid than alfalfa, but thèse cultivar in the proportions of wilted and authors used pea aphid colonies main­ dead seediings between the two repli- tained on faba bean in order to com­ cates of each biomass of aphids (Pr > 0.5 pare a range of host plants and not whateverthe biomass and the cultivar), cultivars of the same host plant. In our data were pooled. Significant différenc­ study, the comparison of the effect of es between the three biomasses were aphids reared either on faba bean or on detected for each cultivar : Pr=3.5 1010 alfalfa revealed no différence in re- for 'CUF 101' and Pr=7 1020 for 'Mil- sponse to aphid infestation between feuil'. The infestations using the low- susceptible ('Milfeuil') and résistant est biomass of aphids (180 mg) were ('CUF 101 ') cultivars when aphids were not sufficient to induce a satisfying lev- reared on faba bean. It is likely due to el of wilting and mortality of seediings the fact that aphids reared on faba bean (Table 3) and to detect différences be­ were not conditioned to their new host tween the résistant and sensible culti- (alfalfa) and resulting in a higher mor­ vars ; we were unsuccessfull in reach- tality of thèse aphids. Thèse results are ing 60% of dead and wilted seediings in accordance with those mentioned by for 'Milfeuil1. The biomasses of 360 mg Frazer (1972) collecting PA from four and 540 mg were équivalent whatever species (broom, broad bean, white clo- the cultivar. They induced 26 to 28% of ver and alfalfa). He noticed that broad wilted plus dead seediings for'CUF 101' bean was the only plant which allows and 59 to 66% for 'Milfeuil'. Such pro­ the multiplication of the différent strains, portions allowthe distinction between but the pea aphid reared on bean strains résistant and susceptible cultivars. did not succeed on alfalfa 'Ladak'. The Thèse results led us to sprinkle aphids problem of preconditioning has also twice over seediings in an experiment, been emphasized by Schotzko and with 4 or 5 d interval between sprin- Smith (1991), who reported that the klings and to use biomasses of at least fecundity of the Russian wheat aphid, 360 mg per unit of 54.

Table 3. Estimations of proportions p and their standard errors SE(p) of dead plus wilted seediings after infestations with différent aphid biomasses 'CUF 101' 'Milfeuil' Biomass of aphids by infestation Biomass of aphids by infestation (mg unit1) (mg unit1)

180 360 540 180 360 540 TOPROTECTI O P 0.01 0.26 0.28 0.10 0.66 0.59 PH Y SE(p) 0.01 0.00 0.06 0.04 0.03 0.02

144 GIROUSSE ET AL. : RESISTANCE TEST OF ALFALFA TO THE PEA APHID

Diuraphis noxla [Homoptera : Aphid- a low level of infestation (180 mg per idae] (Mordvilko), is influenced by unit) was not sufficient to induce a high the preconditioning effect of wheat mortality of 'Milfeuil' seedlings ; on the 'Stephens', when comparing this culti- contrary, the biomasses of 360 mg and var with oat 'Boarder'. To avoid the delay 540 mg per unitwere équivalent in terms of the aphid effect on alfalfa seedlings of wilted plus dead seedlings. In our during a résistance test, wesuggest that expérimental conditions, 360 mg per PA should be reared on a susceptible unit seemed to be a good compromise cultivar of alfalfa. In our study, we chose between aphid mass production and a to rear aphids on 'Milfeuil'. good level of mortality in 'Milfeuil1 seed­ lings (about 60%). But this aphid biom- Exp. 2. The seedling tests of résis­ ass has to be adjusted to fit two other tance of alfalfa to A. pisum, T. maculata parameters : the phenological stage of and A. kondoï are usually carried on seedlings and the size of expérimental under free-choice conditions, either in unit. The older the seedlings are when the initial tests in the USA (Ortman et infested, the larger the aphid biomass al. 1960), or in the more récent tests in has to be in order to induce a high Australia (Ridland and Berg 1981). This mortality in the susceptible control. gives the advantage of testing antix- Then, to limit the amount of aphids, it enosis and tolérance components of seems that the plants need to be infest­ résistance. Under agronomie condi­ ed as early as possible, that is when the tions, antixenosis is linked to alatae cotylédons are just opened. With re­ choice as reported by Holtkamp and gard to the size of expérimental unit, Clift (1993). Under controlled conditions, with the same number of plants, a larg­ the sélective production of alatae in er unit needs greater amounts of infest- large amounts is rather complicated, or ing aphids than a smaller one ; actually, even impossible. Our mass inoculum of in a large unit, aphids tend to wander. aphids included mainly apterae. More- To prevent this aphid behaviour, it over, we proved the homogeneity of seems to be more convenient to use the response of the cultivars tested in small expérimental units, with equal choice or non-choice experiments. Con- number of plants. Lastly, it is important sequently, we express the view that the to make sure that the quality of infes­ non-choice test is adapted to our aim. tation remains the same in ail experi­ It prevents any mixing of the seedlings ments ; a way to control infestation owing to the séparation of the différent quality is to weigh wingless adults. For batches. instance, in our conditions, if average Exp. 3. The détermination of the level weight is less than 1.5 mg in our con­ of infestation is of primary importance ditions, it would be better to cancel the for the success of the seedling test. test. Otherwise this could lead to a Formerly, it has been described as misinterpretation of the results, conclud- "massive infestation" by Ortman et al. ing the résistance of a cv instead of a (1960), sprinkling additional aphids in lack of aphid effect. orderto maintain population as high as possible. Berberet et al. (1991) noted Repeatability of the test and that the infestation procédure needs a number of replîcates for minimum of two aphids per seedling treatment comparison and additional supplies if necessary. The accuracy of the US seedling test is Other publications refered to the vol­ not reported in literature. It would be ume of aphids by référence to the unit partly due to the difficulty of controlling of infestation (Hackerott et al. 1963; the variability either of aphids, environ- Nielson and Lehman 1977; Ridland and ments or plants that are many causes of Berg 1981). The crucial point is to find variation. The test previously described a good compromise between an easy was conducted for 3 yr and it is inter- mass production of aphids and infesta­ esting to take advantage of data collect- tion levels inducing a suitable rate of ed during thèse independent tests for the mortality in the susceptible cultivar evaluating the repeatability of the test. control. Our results demonstrated that We considered the results obtained for

145 the same line 'Milfeuil' during three comparison of two percentages, hav- independent tests (1993, 1994, 1995) ing 1-G chances in 100 to detect an using five (yr 1993) or six units (yr 1994 absolute différence d between thèse two and 1995). The conditions of infesta- percentages. The number of replicates tions were those described above ; the \n) could be calculated (Cochran 1977; end of infestation occurred when more Dagnelie 1986) using y, the number of than 60% of the 'Milfeuil' seediings wiited plus dead seediings among the were wiited or dead. The estimation of 54 seediings of each unit : the proportions of wiited plus dead seediings at the end of infestation for 2 2 each of the three tests are reported on 2(u1_ + "J * [5] Table 4. There was no différence be- n = tween the proportions obtained inde- pendently in the conditions of our growth chamber. Therefore, trying to where n is the number of replicates control as many as possible of thèse (units), d the absolute différence on parameters leads to results that clearly y to detect a différence between the prove the homogeneity of responses of two treatments (defined in number of a seedling test conducted in growth wiited plus dead seediings ; for exam­ chamber on alfalfa. ple, d=5.4 corresponds to an absolute différence of 10% between the perfor­ Using this test, significant différenc­ mance of the 2 treatments), a2 the es between cultivars could be difficult variance of y, a and (3 respectively the to detect if the résistance levels of the Type 1 and Type 2 error and u the cultivars are too close. It dépends on quantile of the Normal distribution. The the number of plants and/or units that variability between units in the tests are tested but the cost of sampling could conducted from 1993 to 1995 on 'Mil­ then become high to obtain a small gain feuil' was from 3.92 to 5.32 (standard in précision. We can use the results of error of y in Table 4). The number of the same three independent tests de­ replicates to use in order to compare scribed above and the variability of the two proportions of wiited plus dead proportions forthe establishment of the seediings was calculated using the number of replicates necessary for the highest standard error of the tests, i.e.

Table 4. Proportions p and standard errors SE(p) of the proportions of dead plus wiited seediings ; mean number of dead plus wiited seediings per unit (y) and standard errors of y( cr) m three i nde penden t testi

Year N umber of P SE(p) y cr units

1993 5 0.637 0.044 36.4 5.32 1994 6 0.688 0.030 37.2 3.92 1995 6 0.636 0.034 34.3 4.50

Table 5. Number of units (n) to detect dp absolute différence on proportions p or dy absolute différence on dead plus wiited seedling numbers (y), with two first [a) and second (|3) risk levels

a=0.05 ; 0=0.1 a=0.1 ; p=0-2

dp 0.10 0.20 0.30 0.10 0.20 0.30 dy 5.4 10.8 16.2 5.4 10.8 16.2 n 21 6 3 12 3 2

146 GIROUSSE ET AL. : RESISTANCE TEST OF ALFALFA TO THE PEA APHID

CT=5.32. Results are presented on Table Bournoville, R. 1976. Enquête sur les insect­ 5, using for the calculations two first (a) es nuisibles à la luzerne en production de and second type ((3) errors, and three fourrage Pages 83-86 in : Compte-rendus absolute différences to detect between réunion EUCARPIA, Groupe Medicago sativa, VURV (eds.) Piestany Tchécoslova­ the two proportions. When the best risk quie. levels are required, we could detect an Bournoville, R. 1980. Varietal characteristic absolute différence of 10% between two under French conditions of some alfalfa treatments by using 21 units of 54 seed- cultivars selected for résistance to two lings per treatment. This leads to a very insects. EPPO Bull. 10 : 317-322. powerful design. For instance, in the Bournoville, R., and B. Comte. 1977. Pre­ context of plant breeding i.e. when miers résultats sur l'étude en France du comparing and screening cultivars for comportement varietal de la luzerne PA résistance, the use of six units per [Medicago sativa L.) à l'égard du puceron du pois {Acyrthosiphon pisum H.) (Ho- cultivar or treatment could be sufficient moptera : Aphididae) et du phytonome to detect différences of 20 or 30% in (Hypera variabilis H.) (Coleoptera : Cur- cultivar performance. Even more, if the culionidae). Bull. SROP-OILB 3 : 97-101. purpose is to detect highly résistant or Carnahan H.L., R.N. Peaden, F.V. Lieberman, susceptible cultivars, two units could and R.K. Petersen. 1963. Differential re­ be sufficient. This validâtes the use of actions of alfalfa varieties and sélections two replicates per treatment in experi- to the pea aphid. Crop Sci. 3 : 219-222. ments 1, 2, 3. It is obvious that if the Cochran, W.G. 1977. Sampling techniques, variability between units is higher, due John Wiley and Sons, N.Y., 428 pp. Dagnelie, P. 1986. Théorie et méthodes statis­ to différent expérimental conditions or tiques. Les presses agronomiques de device, or to the cultivars, the number Gembloux, Belgique, 463 pp. of replicates required to detect the same Ellsbury, M.M., and M.W. Nielson. 1981. différence or to obtain the same level of Comparative host plant range studies of précision will hâve to be increased. On the blue alfalfa aphid, Acyrthosiphon the other hand, the number of repli­ /Conc/o/Shinji, and the pea aphid, Acyrtho­ cates has been calculated using units of siphon pisum (Harris) (Homoptera : Aphi­ 54 seedlings. Changing the number of didae). Tech. Bull., USDA 1639 : 1-14. seedlings per unit is équivalent to Frazer, B.D. 1972. Population dynamics and récognition of biotypes in the pea aphid changing the test and requires to re- (Homoptera : Aphididae). Can. Entomol. establish the number of replicates. 104 : 1729-1733. Girousse, C, and R. Bournoville. 1994. Bi- ological criteria of the pea aphid Acyrtho­ ACKNOWLEDGEMENTS siphon pisum Harris and varietal résis­ tance of Alfalfa. Eucarpia/FAO meeting "Management and breeding of perennial The authors are indebted to Dr. R. lucern for diversified purposes", 4-8 Sept. Michaud for his helpful comments on 1994, Lusignan, France, 251-253. the manuscript. We wish to thank Lau­ Hackerott, H.L., E.L. Sorensen, T.L. Harvey, rence Favre and Marilyn Vandier for E.E. Ortman, and R.H. Painter. 1963. Re­ technical assistance. This programme actions of alfalfa varieties to pea aphids was partly supported by a grant provid- in the field and greenhouse. Crop Sci. 3 : ed by a plant breeders association : 298-301. l'ACVF (Association des Créateurs de Holtkamp, R.H., and A.D. Clift. 1993. Estab­ lishment of three speciesof alfalfa aphids Variétés Fourragères). on 24 cultivars of alfalfa. Austr. J. Agric. Res. 44 : 53-58. Isaak, A., E.L. Sorensen, and E.E. Ortman. REFERENCES 1963. Influence of température and hu- midity on résistance in alfalfa to the spot- Berberet, R.C., J.L. Caddel, and A.A. Zarra- ted alfalfa aphid and pea aphid. J. Econ. bi. 1991. Pea aphid résistance. In Stan­ Entomol. 56 : 53-57. dard tests to characterize alfalfa cultivars. Isaak, A., E.L. Sorensen, and R.H. Painter. C.C. Fox, R. Berberet, F.A.A. Gray, C.R. 1965. Stability of résistance to pea aphid Grau, D.L. Jessen, and M.A. Peterson and spotted alfalfa aphid in several alfal­ (eds.), North American Alfalfa Improve- fa clones under various température ré­ ment Conférence, USDA/ARS, Beltsville, gimes. J. Econ. Entomol. 58 : 140-143. MD, I.4.

147 Kugler, J.L., and R.H. Ratcliffe. 1983. Résis­ Smith, O.F., and R.N. Peaden. 1960. A meth- tance in alfalfa to a red form of the pea od of testing alfalfa plants for résistance aphid (Homoptera : Aphididae) J. Econ. to the pea aphid. Agron. J. 52 : 609-610. Entomol. 76 : 74-76. Smith, CM., Z.R. Khan, and M.D. Pathak. Nielson, M.W., and W.F. Lehman. 1977. 1994. Techniques for evaluating insect Multiple aphid résistance in CUF 101 al­ résistance in crop plants. CRC Press, Boca falfa. J. Econ. Entomol. 70 : 13-14. Raton, Florida, 320 pp. Ortman, E.E., E.L. Sorensen, R.H. Painter, Sorensen, E.L., R.A. Byers, and E.K. Horber. T.L. Harvey, and H.L. Hackerott. 1960. 1988. Breeding for insect résistance. Sélection and évaluation of pea aphid Pages 859-902 in Alfalfa and Alfalfa résistant alfalfa plants. J. Econ. Entomol. Improvement, A.A. Hanson et al. (eds)., 53 : 881-887. American Soc. Agronomy, Madison, Ridiand, P.M., and G.N. Berg. 1981. Seediing Wisconsin. résistance to spotted alfalfa aphid of al­ Summers, CG. 1988. Cultivar and tempér­ falfa and annual medic species in Victo­ ature influence on development, surviv- ria. Aust. J. Exp. Agric. Anim. Husb. 21 : al, and fecundity in four successive 59-62. générations of Acyrthosiphon kondoi(Ho­ Schalk, J.M., S.D. Kindler, and G.R. Man- moptera : Aphididae). J. Econ. Entomol. glitz. 1969. Température and the préfér­ 81 : 515-521. ence of the spotted alfalfa aphid for ré­ sistant and susceptible alfalfa plants. J. Econ. Entomol. 62 : 1000-1003. Schotzko, D.J, and CM. Smith. 1991. Effects of preconditioning host plants on popu­ lation development of russian wheat aphids (Homoptera : Aphididae). J. Econ. Entomol. 84 : 1083-1087.

148