Bulletin of Entomological Research (2000) 90, 233–244 233

Host-correlated morphological variationof Myzus persicae (Hemiptera: Aphididae)

populations in Greece

J.T. Margaritopoulos1, J.A. Tsitsipis1*, E. Zintzaras1 andR.L. Blackman2

1Laboratory of Entomology and Agricultural Zoology, Department ofAgriculture, Crop and Animal Production, University of Thessaly, Pedion

Areos 38334, Volos, Greece: 2Department of Entomology, The NaturalHistory Museum, Cromwell Road, London, SW7 5BD, UK

Abstract

Morphological variation in nine characters of 157 clones of Myzus persicae(Sulzer) was examined by multivariate analysis. The clones were collected frompeach, Prunus persica, the primary host and the secondary hosts tobacco, Nicotianatabacum, cabbage, Brassica oleracea, sugarbeet, Beta vulgaris and pepper Capsicumannuum. The 156 clones originated from various regions of Greece, both in thenorth, where a large part of the population has an annual bisexual generation onpeach, and in more southerly regions, where populations are predominantlyunisexual. One clone was collected from tobacco in Caserta, Italy. All clones werelaboratory-reared on potato. Canonical variate analysis, hierarchical clusteranalysis and a non-parametric classification tree method both revealedmorphological differences associated with the host-plant on which they werecollected. The scores of the first two canonical variates separated the tobacco-feeding clones from those originating from other secondary host-plants. However,in tobacco-growing areas the tobacco-feeding form predominated in springpopulations on peach, and was sometimes found on other secondary hosts. Inaddition, using cluster analysis, the clones from tobacco which were sampled inthe most southeasterly region showed a relatively large phenotypic distance fromthose collected further north and west. Moreover, clonal phenotypes were affectedboth by host plant and by long-term parthenogenetic rearing. However, in spite ofthese effects, the tobacco form was generally distinguishable from aphidsoriginating from other hosts, indicating that the difference must have a geneticbasis. In separate analyses of the clones originating from secondary hosts noassociation was found between morphology and either life cycle category orcolour. Discriminant analysis showed that 89% of 1723 specimens could becorrectly classified into the two groups.

Introduction

In 1987, Blackman carried out morphometric studies onnumerous samples of the Myzus persicae group (Hemiptera:Aphididae), from different host-plants from four continents.

He showed that samples originating from tobacco, Nicotianatabacum L. (Solanaceae), could be distinguished using amultivariate approach. He suggested that the differentmorphological traits of the tobacco form were well stabilizedand under a polygenic control. In the same study, Blackmanprovided evidence supporting the genetic isolation of thetobacco form. The tobacco-feeding populations werethought to be anholocyclic, and therefore unable to producehybrids with Myzus persicae (Sulzer). On the basis of these

* Author for correspondenceFax: +30 421 61957E-mail: tsitsipi@uth.gr

data, the tobacco-feeding form was described as a newspecies Myzus nicotianae Blackman (Blackman, 1987).

However, the tobacco form is now known to showconsiderable variation in its life cycle. In many tobacco-growing areas of the world, sexual reproduction is probablyentirely absent (Blackman & Spence, 1992). On the otherhand, in Japan and Central Asia there occur holocyclic andandrocyclic populations of the M. persicae group feeding ontobacco (Zagorovskii, 1947; Kolesova et al., 1980; Takada,1986). In Greece, tobacco-feeding clones were found tomigrate from peach, Prunus persica L. (Rosaceae), to tobacco(Blackman & Spence, 1992). In addition, in northern Greecearound the main peach-growing regions, a high proportionof clones of M. persicae s.l. feeding on tobacco and on otherhosts are holocyclic (Margaritopoulos & Tsitsipis,unpublished data).

Blackman & Spence (1992) found a polymorphism in theenzyme glutamate oxoacetate transminase (GOT) thatenables discrimination of the tobacco feeding form with upto 90% reliability. Myzus persicae is monomorphic for GOT,whereas M. nicotianae is polymorphic, with two forms ofslightly different mobility. Nevertheless, DNA sequences offragments of the E4 and FE4 genes, involved in esterase-based mechanisms of insecticide resistance, have been foundto be identical between M. persicae and M. nicotianae (Field etal., 1994). The results of a recent RAPD analysis of severalclones of the two groups failed to detect any consistentdifferences in RAPD banding pattern between M. persicae s.s.and M. nicotianae (Margaritopoulos et al., 1998). Therefore,there is doubt about whether M. nicotianae should beregarded as a distinct species. However, using the RAPDmethod, a single-band difference was found significantlyassociated with host plant, which contributes to the idea thatthere may be a host race on tobacco.

The present study is aimed at examining the pattern andthe extent of morphological variation amongst clones of M.persicae originating from tobacco and non-tobacco hosts fromvarious regions of Greece. In addition, the effects of host-plant and long term parthenogenetic rearing on aphidmorphology and the association between morphologicaldifferences and colour or life cycle category were alsoexamined.

Materials and methods

During the years 1995–1997, 156 samples of M. persicaewere collected from tobacco, cabbage, Brassica oleracea L.(Brassicaceae), sugarbeet, Beta vulgaris L. (Chenopodiaceae),pepper Capsicum annuum L. (Solanaceae) and peach fromdifferent regions of Greece. In addition, one cloneoriginating from tobacco was collected from Caserta in Italyin 1998. Most of the clones originating from peach and allthe clones from pepper were collected away from tobacco-growing regions. On the other hand, clones from cabbage,sugarbeet and some clones from peach were collected fromtobacco-growing regions (fig. 1).

A large c. 1 ha field was sampled from each region eachyear. One sample was collected per plant and it consisted ofone to three leaves infested by aphids. In annual crops,plants were sampled every four to five rows and at leastevery 5 m along the row. On peach, samples were takenevery four to five trees along the row. Each sample was putin a self-sealing plastic bag, slightly inflated and containingone piece of paper towel. Bags were placed in isolated

plastic containers containing ice packs and transferred to thelaboratory. Clonal cultures were established from a singleapterous adult virginopara from each sample, except for twoclones from peach that were started from alatae. Dataconcerning the origin of each clone, its colour, theirbiological cycle and host plant are given in table 1. Mostclones were reared for 1–3 generations on excised potatoleaves in plastic boxes (7.7 cm 3 4.5 cm 3 2 cm) (Blackman,1971) in a growth chamber at 17°C with a photoperiod of 16 h light per day. Four clones collected from peach intobacco-growing regions (Aridea, Alexandria) were rearedfor 7–8 months under the same conditions. Approximately10 (min: 4, max: 20) apterous adult virginoparae from eachclone were preserved in tubes filled with one part lactic acid(75% w/w): two parts ethyl alcohol (95%), until slidepreparation. Aphids were mounted on slides according tothe method of Blackman & Eastop (1984).

The aphids included in the study comprised 157 clonalcultures; 36 cultures (323 specimens) from peach, 95 (819specimens) from tobacco, 15 (155 specimens) from pepper,nine (101 specimens) from cabbage and two from sugarbeet(20 specimens). Three clones which were collected fromtobacco, two from the same population from Anavra (codes:T1, T2) and one from Alexandria (code: T3) were alsomeasured after rearing for two years on potato. Moreover,two clones originating from peach at Lehonia (code: L6) andPortaria (code: P) were also measured one and three yearsrespectively after collection from peach. The clones werereared throughout this period on excised potato leaves. Toexamine the short-term effect of host-plant on morphology,four of the former clones (T1–T3, P) were also reared for onegeneration on cabbage, broad bean Vicia faba L. (Fabaceae)and three on tobacco (clones T1–T3).

Nine characters were measured according to Ilharco &van Harten (1987). The same ones that Blackman (1987)found to give satisfactory separation between the aphidsfrom tobacco and those from other hosts. These were: lengthof third antennal segment (ant III), length of base of sixthantennal segment (base VI), length of terminal process ofsixth antennal segment (pt), length of ultimate rostralsegment (urs), length of hind femur (hf), length of secondsegment of hind tarsus (ht2), length of siphunculus (ls),maximal width of distal swollen part of siphunculus (mws)and length of cauda (lc). All measurements were carried outunder a phase contrast microscope (Leica DRMB).

To examine the separation of the tobacco form, the dataof the 157 clones were submitted to a canonical variateanalyses (CVA) (Krzanowski, 1990). Each clone was treatedas one group. The CVA has been proved to be a powerfultool for the discrimination of taxa in aphids when the clonesare used as groups in the analysis (Krzanowski, 1990;Blackman, 1992). To examine the effect of host-plant onaphid morphology, a series of canonical variate analyses wasrun comparing only the data from the three tobacco clones(T1–T3) and the one peach clone (P), which were reared onvarious hosts. Different sets of clones, listed in table 1, werecompared to examine variation in morphological charactersbetween clones differing in life cycle category and colour.

The separation of the tobacco form was furtherinvestigated by applying a classification tree method(Zintzaras et al., 1994) on the individual aphids originatingfrom different host-plants. The latter is a non-parametricdiscriminant method which uses as splitting criterion theincrease in purity when a node is split into two subnodes,

234 J.T. Margaritopoulos et al.

and the size of the tree is controlled by a threshold level forthe apparent misclassification rate (AMR) of the tree aftereach splitting step.

The morphological variation between groups of clonescollected from different hosts and regions was examinedusing single linkage hierarchical cluster analysis (Sneath &Sokal, 1973) based on the Mahalanobis distances.Hierarchical cluster analysis was also used to examine themorphological variation between groups of tobacco clonesoriginating from different regions. However, the two clonesfrom Preveza and Caserta (Italy) were excluded from thisanalysis, since it is difficult to characterize the population ofa region from a single clone.

The separation of the tobacco form from the non-tobaccoform at the level of individual aphids was examined using

the Fisher‘s linear discriminant function (LDF) (Krzanowski,1990).

No transformation was done on data before analysis. Thestatistical analyses were performed on the untransformeddata using SPSS v.6.1 and Statistica v.4.3.

Results

The mean lengths of nine characters for M. persicae specimensoriginating from various plants are shown in table 2. In thefirst canonical variate analysis all 157 clones were examinedtogether. Figure 2a is a plot of the mean scores on the firstand second canonical variate (CV), which togetheraccounted for 47% of the total variation in the data, and fig.2b is a plot of CV1 and CV3, which together accounted for

Morphological variation of M. persicae populations 235

GREECE

Scale 1: 5000000

PELOPONNESE

9 Pep

CRETE

FYROM

BULGARIA

THRACE

ALBANIA

11T1 P MACEDONIA

2 T, P, C, Sb

10 T3 T

CENTRALGREECE

6 T12 T

7 T8 T

4 P5 P, Pep

Fig. 1. Collection sites and host-plants of the clones of Myzus persicae in Greece. 1, Aridea; 2, Alexandria; 3, Tsaritsani; 4, Larisa; 5, Volos;6, Anavra; 7, Agrinio; 8, Amphiklia; 9, Heraklio; 10, Katerini; 11, Kilkis; 12, Preveza; Fyrom, Former Yugoslavian Republic of Macedonia(T, tobacco; P, peach; Pep, pepper; C, cabbage; Sb, sugarbeet).

39% of the total variation. Most of the clones from tobacco,and eight out of nine clones originating from peach withintobacco-growing areas (Alexandria, Aridea), formed adifferent cluster from the clones originating from other host-plants. The four clones originating from peach in theAlexandria/Aridea region, which were measured after eightmonths of rearing in the laboratory on excised potato leaves,still clustered with the clones from tobacco, with highervalues of CV1 than the clones measured one to threegenerations after collection from peach (figs. 2a,b).

The separation between clones from tobacco and clonesfrom other hosts was almost entirely due to CV1, whichaccounted for 26% of the variation in the data, but was notcomplete. For example, two clones from cabbage at

Alexandria were clustered with the aphids from tobacco.However two clones from peach, one from Volos (Lehonia)and one from Larisa (Falani), were grouped with a fewclones from tobacco outside the main cluster of aphids fromtobacco. In addition, four green clones (three holocyclic andone anholocyclic) from tobacco were located near the clonesfrom pepper and cabbage.

The character with the highest correlation to CV1 was theultimate rostral segment (urs). The highest values of CV1 forclones other than those from tobacco were recorded mainlyin clones from cabbage. Most of the clones from pepper hadhigher values on CV2, and lower values on CV3, than clonesfrom other non-tobacco crops.

In separate analyses of the clones originating from

236 J.T. Margaritopoulos et al.

Table 1. Clones of Myzus persicae used in the study.

Region Host-plant Sampling year Holocyclic1/ Green/redexamined clones

clones1995 1996 1997

no. date no. date no. date

Lehonia, Volos NT3 peach 5 25 May 13 23 May 4 23 May 22/23 23/0Falani, Larisa NT peach – – – – 2 22 May 2/2 2/0Portaria, Volos NT peach 3 22 May – – – – 3/3 3/0Aridea T4 peach – – – – 3 15 Jun 3/3 0/3Alexandria T peach – – – – 6 11 Jun 6/6 6/0Velestino, Volos NT pepper 2 20 Aug 4 27 May 1 10 Jun 2/7 6/1Dimini, Volos NT pepper – – 4 15 Jun – – 4/4 4/4Heraklion, Crete NT pepper – – 4 27 Jul – – 0/4 4/0Alexandria T cabbage – – 9 12 Jun – – 5/9 9/0Alexandria T sugarbeet – – 2 12 Jun – – 2/2 1/1Agrinio T tobacco 7 15 Sep 4 10 Jul – – 2/11 6/5Amphiklia T tobacco – – 7 2 Jul – – 1/7 0/7Anavra T tobacco 8 3 Sep 10 28 May,14 Jun – – 0/18 8/10Tsaritsani T tobacco 7 5 Sep 3 30 Jul – – 3/10 6/4Preveza T tobacco – – – – 1 30 Jun 0/1 0/1Katerini T tobacco – – – – 18 4 Jun, 9 Sep 14/18 14/4Alexandria T tobacco – – 10 12 Jun 2 11 Jun 5/12 6/6Kilkis T tobacco 17 5 Sep – – – – 9/17 16/1Caserta (Italy)2 T tobacco – – – – 1 7 Jul 0/1 0/1

1The other clones are anholocyclic or androcyclic. 2Collected in 1998. 3NT, Non-tobacco growing region. 4Tobacco growing region.

Table 2. Mean lengths (±SE) in mm of nine characters of apterae of Myzus persicae collected from different hosts and regions of Greece.

Host No. individuals ant III base IV pt urs hf ht2 ls mws lc

Peach2 238 0.408 0.130 0.458 0.116 0.589 0.113 0.474 0.051 0.202(0.003) (0.006) (0.002) (0.0003) (0.004) (0.005) (0.003) (0.0003) (0.001)

Pepper 155 0.458 0.133 0.476 0.118 0.649 0.117 0.533 0.052 0.231(0.003) (0.009) (0.003) (0.0004) (0.005) (0.006) (0.004) (0.0004) (0.002)

Cabbage 101 0.412 0.133 0.480 0.118 0.612 0.113 0.504 0.052 0.217 (0.004) (0.01) (0.004) (0.0007) (0.006) (0.006) (0.005) (0.0005) (0.002)

Sugarbeet 20 0.432 0.134 0.476 0.118 0.637 0.113 0.522 0.058 0.232 (0.006) (0.019) (0.004) (0.0007) (0.009) (0.01) (0.008) (0.001) (0.002)

Tobacco 819 0.430 0.136 0.505 0.124 0.630 0.113 0.515 0.051 0.221(0.002) (0.003) (0.001) (0.0002) (0.002) (0.003) (0.002) (0.0002) (0.001)

Peach1 85 0.435 0.145 0.496 0.122 0.604 0.112 0.499 0.051 0.217(0.005) (0.014) (0.005) (0.0006) (0.01) (0.016) (0.007) (0.0005) (0.002)

1Clones collected from peaches in tobacco-growing regions.2Clones collected from peaches away from tobacco-growing regions.ant III, length of third antennal segment; base VI, length of base of sixth antennal segment; pt, length of terminal process of sixthantennal segment; urs, length of ultimate rostral segment; hf, length of hind femur; ht2, length of second segment of hind tarsus; ls,length of siphunculus; mws, maximal width of distal swollen part of siphunculus; lc, length of cauda.

Morphological variation of M. persicae populations 237

Fig. 2. Plots of the mean scores of the first three canonical variates for 157 clones of Myzus persicae collected from different regions ofGreece. +, Holocyclic clones from tobacco; e, anholocyclic clones from tobacco; j, clones collected from cabbage; 3, clones collectedfrom sugarbeet; ., clones collected from pepper; s, clones collected from peach away from tobacco-growing regions; m, clones collectedfrom peach in tobacco-growing regions. 168, clone collected from Caserta, Italy; 164, 165, 166, 167, clones collected from peach intobacco-growing regions which were measured 8 months after the collection.

secondary hosts no association was found betweenmorphology and either life cycle category or colour (plots ofthe canonical variates are not shown). There are probablymultiple genotypes of each colour and life cycle category.

To investigate the effect of host-plant on aphidmorphology, three clones from tobacco (T1–T3) and onefrom peach (P; away from tobacco-growing area) werereared for one generation on different host-plants and thencompared by CVA. In this analysis (fig. 3), CV1 scores werehighly correlated with a size index based on the sum of thenine characters (r = 93.2), and it is clear that most of thevariation of CV1 (49% of the total variation) is due to thehost plant on which the clones were reared (fig. 3a). Forexample, specimens of all clones reared on cabbage weresignificantly smaller than when reared on other plants (table3). After removing variation due to CV1, scores on CV2 andCV3 separated the clones from tobacco from that collectedon peach away from the tobacco-growing region, regardlessof the plants on which they were reared (fig. 3b).

The next step of analysis was to examine how clonesT1–T3, P, and clone L6 (samples of which were measuredone to three generations after collection from their host-plantand one to three years later) clustered in relation to the other157 clones. Figure 4 is a plot of the first two canonicalvariates, which together account for 49% of the totalvariation of data. Clones T1–T3, regardless of the plant onwhich were reared, were grouped with the aphids fromtobacco, as were the specimens measured two yearspreviously. The specimens of clones P and L6, that werereared on various host-plants (only potato for L6) after threeand one year respectively of rearing in the laboratory onexcised potato leaves, remained distinct from the tobaccoclones. Comparing the samples of the five clones, measuredafter short-term (one to three generations) or long-term (oneor more years) rearing on potato, some changes in value ofthe first two CVs were observed, especially CV1, and thesewere more apparent in three clones (T1, T2, P). Significantincreases in the mean lengths of almost all the characters inthese three clones was observed, but only for threecharacters in clone L6, and in T3 the mean length of somecharacters decreased (table 4). However, irrespective of the

rearing time and the plant on which they were reared, cloneswere still located in their original groups (aphids on tobaccovs. non-tobacco aphids) (fig. 4).

The classification tree method shows (fig. 5) that theaphids from tobacco were either mainly separated fromother individuals (nodes 8 and 11), or mixed with aphidsoriginating from peach in tobacco-growing regions (node 3).However, the separation was not complete, as in canonicalvariate analysis, since some individuals from peach awayfrom tobacco-growing regions and from pepper were mixedwith the aphids from tobacco in node 3, this might be due tothe high variability of these individuals. The majority of thepeach aphids (away from tobacco-growing regions) aremixed with the pepper and the cabbage aphids (node 6),one quarter of them separating at node 10. A largeproportion (one third) of the pepper aphids was segregatedat node 5. The individuals of one clone from peach (neartobacco-growing areas) were located in the same node withaphids from peach away from the tobacco-growing regions(node 6). However, this clone was not classified as thetobacco form based on canonical variate analysis.Individuals of the two clones from cabbage, which wereclassified as the tobacco form by canonical variate analysis,were mixed with aphids from tobacco in the classificationtree (node 3).

Figure 6a shows the dendrogram of cluster analysisbased on data of Mahalanobis distances (table 5) betweenclones that were collected from different host-plants andregions. Two main clusters stand out in the tree. The oneconsisted of clones originating from peach and pepper faraway from tobacco growing regions and from cabbage andsugarbeet near tobacco fields. The second cluster consistedof clones that were collected from tobacco and peach(Alexandria and Aridea regions) close to tobacco fields.However, the tobacco clone from Caserta was separatedfrom the two former clusters. In addition, fig. 6b shows theresults of cluster analysis that was applied to the clonesoriginating from tobacco in different regions (table 6). Theseven clones from Amphiklia showed a relatively largephenotypic distance from those collected from central andnorth Greece.

238 J.T. Margaritopoulos et al.

Fig. 3. Plots of the scores of the first three canonical variates of four clones of Myzus persicae reared on different host-plants for onegeneration. T1 and T2, red androcyclic clones collected from tobacco from Anavra region; T3, green holocyclic clones collected fromtobacco from Alexandria region; P, green holocyclic clone collected from peach away from tobacco-growing regions. +, tobacco; u,cabbage; ., broad bean; d, potato.

A final step of analysis was to examine the separation ofthe tobacco form at the level of individual aphids using theFisher‘s linear discriminant function (LDF). The total correctclassification score for all the examined specimens was 89%.However, classification results for the tobacco form and forthe aphids originating from other hosts were 92 and 83%respectively. The character of primary importance indistinguishing between the two groups was the ultimaterostral segment (table 7).

Discussion

In the present study, the morphological variation foundin clones of M. persicae s.l. was correlated with the host plantfrom which they were collected. Using a multivariateapproach, the clones originating from tobacco, regardless oflife cycle category, colour and sampling region, weregenerally separated from those feeding on other secondaryhosts. The results obtained by canonical variate analysesusing the clones as a grouping factor were also supported bythe results obtained using the classification tree method onindividual aphids and the hierarchical cluster analysis usingthe collection region as grouping factor. The results of thepresent study agree with those reported by Blackman (1987)and Blackman & Spence (1992), who established the discrete

morphology of the tobacco form after examining samplesfrom various countries including Greece.

Most clones found on secondary host-plants in a tobacco-growing region (Alexandria) were not of the tobacco form.However, the CV1 scores of two clones feeding on cabbageconformed to those of tobacco-feeding ones, indicating thatthe tobacco form is able to accept and colonize other hosts.However, aphids from tobacco were not sampled from otherhost-plants in regions where tobacco is not common.Probably there is a geographical distribution of the tobaccoform associated with the presence or absence of its mainhost-plant. The fact that both forms were found in tobacco-growing regions on both peach and secondary hostssupports the statement of Blackman & Spence (1992), thatthe two taxa can keep their host-related differences due to amechanism of partial or total reproductive isolation.Additional evidence for this is the single RAPD banddifference between clones collected from tobacco and otherhost-plants (Margaritopoulos et al., 1998).

It is well known that the morphological characters inaphids are affected not only by genotype (Wool, 1977) but byenvironmental components, such as the species (Moran,1986, 1988; Gillham & Claridge, 1994) and physiologicalcondition of host plant, and temperature (Blackman &Spence, 1994; Dixon, 1998). The question that arises often in

Morphological variation of M. persicae populations 239

Table 3. Mean lengths (±SE) in mm of nine characters of clones of Myzus persicae reared on different host-plants.

Clones Host-plants ant III* base IV pt urs hf ht2 ls mws lc

T1 potato (n=20) 0.460 b 0.132 ab 0.499 ab 0.126 a 0.692 ab 0.117 ab 0.557 b 0.057 b0.239 b(0.005) (0.002) (0.005) (0.001) (0.006) (0.001) (0.005) (0.001) (0.002)

cabbage (n=20) 0.423 a 0.123 a 0.486 a 0.127 a 0.613 a 0.113 a 0.407 a 0.054 b 0.208 a(0.007) (0.001) (0.004) (0.001) (0.009) (0.001) (0.008) (0.001) (0.004)

broad bean (n=20) 0.476 c 0.135 b 0.518 c 0.126 a 0.715 b 0.120 b 0.559 b 0.053 a 0.239 b(0.005) (0.001) (0.004) (0.001) (0.005) (0.006) (0.004) (0.006) (0.002)

tobacco (n=20) 0.479 c 0.135 b 0.513 bc 0.130 a 0.732 b 0.120 b 0.582 c 0.055 b 0.248 b(0.004) (0.001) (0.006) (0.001) (0.006) (0.006) (0.004) (0.001) (0.003)

T2 potato (n=18) 0.482 b 0.134 a 0.569 a 0.128 a 0.727 c 0.129 c 0.605 b 0.054 0.249 b(0.004) (0.001) (0.004) (0.001) (0.005) (0.001) (0.005) (0.001) (0.003)

cabbage (n=20) 0.382 a 0.131 b 0.512 a 0.127 a 0.573 b 0.112 a 0.456 a 0.052 0.203 a(0.005) (0.007) (0.005) (0.001) (0.006) (0.001) (0.006) (0.001) (0.003)

broad bean (n=20) 0.483 b 0.147 c 0.521 ab 0.127 a 0.694 a 0.118 b 0.562 c 0.052 0.246 b (0.005) (0.001) (0.008) (0.0005) (0.005) (0.001) (0.005) (0.001) (0.003)

tobacco (n=20) 0.475 b 0.143 b 0.550 b 0.127 a 0.738 c 0.123 c 0.609 b 0.054 0.270 c (0.006) (0.002) (0.009) (0.001) (0.009) (0.001) (0.005) (0.0005) (0.003)

T3 potato (n=21) 0.450 c 0.142 a 0.549 0.125 a 0.651 d 0.118 c 0.532 d 0.054 b 0.244 c (0.005) (0.002) (0.006) (0.001) (0.007) (0.001) (0.005) (0.001) (0.003)

cabbage (n=22) 0.369 a 0.138 a 0.515 0.125 a 0.565 a 0.111 a 0.436 a 0.050 a 0.206 a(0.006) (0.001) (0.004) (0.0005) (0.007) (0.001) (0.007) (0.001) (0.003)

broad bean (n=19) 0.417 b 0.139 a 0.513 0.122 b 0.587 b 0.109 ab 0.474 c 0.050 a 0.208 a(0.007) (0.002) (0.006) (0.001) (0.009) (0.001) (0.006) (0.001) (0.004)

tobacco (n=20) 0.424 b 0.139 a 0.498 0.125 a 0.626 c 0.113 b 0.497 b 0.052 ab 0.230 b(0.003) (0.002) (0.028) (0.0005) (0.005) (0.0005) (0.009) (0.0009) (0.002)

P potato (n=20) 0.470 b 0.143 b 0.541 c 0.119 b 0.686 b 0.119 b 0.567 b 0.055 b 0.229 b (0.003) (0.004) (0.004) (0.0006) (0.005) (0.001) (0.003) (0.008) (0.001)

cabbage (n=11) 0.376 a 0.134 a 0.488 a 0.117 a 0.544 a 0.111 a 0.424 a 0.511 a 0.179 a(0.007) (0.005) (0.008) (0.0006) (0.007) (0.001) (0.006) (0.001) (0.003)

broad bean (n=22) 0.476 b 0.141 b 0.528 b 0.119 b 0.680 b 0.117 b 0.554 b 0.545 b 0.224 b (0.007) (0.006) (0.004) (0.0006) (0.01) (0.001) (0.009) (0.0006) (0.003)

T1 and T2: red androcyclic clones collected from tobacco from Anavra region. T3: green holocyclic clone collected from tobacco fromAlexandria region. P: green holocyclic clone collected from peach, away from tobacco growing regions (Portaria, Volos). For each clone mean values of specimens followed by different letter are significantly different (P < 0.05). Duncan‘s test was used.*See Table 2 for explanation of the characters.

240 J.T. Margaritopoulos et al.

T

PT+

C

PT–

Pep

Sb

161

163

162

P 169

L6

173158

T3

160157

156159

155153

170T1

154171

172T2

3

2

1

0

–1

–2

–3

–4–4 –3 –2 –1 0 1 2 3 4 5

CV2

CV

1

Fig. 4. Plot of the mean scores of the first two canonical variates for 157 clones of Myzus persicae, and of samples from some of these clonesafter various experimental treatments. To simplify the diagram, the limits of scatter of the scores for the majority of the clones arerepresented by areas that are shaded according to the host plants on which the clones were collected (see key, and compare fig. 2a), but witha few outliers indicated by symbols (see key). T, clones from tobacco; PT+, clones from peach in tobacco-growing regions; C, clones fromcabbage; PT-, clones from peach away from tobacco-growing regions; Pep, clones from pepper; Sb, clones from sugarbeet. Samples fromlineages reared experimentally are indicated by numbered points. Of the clones collected from peach away from tobacco-growing regions,clone P was also measured after 3 years of laboratory rearing on potato (161) and after one generation on cabbage (162) and broad bean(163), however, clone L6 was also measured after one year of rearing on potato (169). Three clones collected from tobacco, T1, T2 and T3were also measured after 2 years of laboratory rearing on potato (153, 157, 170) and after one generation on cabbage (154, 158, 171), broadbean (155, 159, 172) and tobacco (156, 160, 173).

Table 4. Mean lengths (±SE) in mm of nine characters of Myzus persicae clones collected from tobacco (T1–T3) and peach (P and L6),reared for different periods of time on potato excised leaves.

Clone Rearing time after n ant III* base IV pt urs hf ht2 ls mws lcthe collection

L6 2–3 generations 14 0.396 a 0.124 a 0.471 a 0.116 a 0.561 a 0.110 a 0.459 a 0.049 a 0.187 a(0.004) (0.006) (0.005) (0.001) (0.008) (0.001) (0.008) (0.001) (0.003)

L6 1 year 12 0.410 a 0.135 b 0.470 a 0.117 a 0.561 a 0.110 a 0.487b 0.055 b 0.200 b(0.007) (0.006) (0.01) (0.001) (0.01) (0.001) (0.008) (0.002) (0.004)

P 1–3 generations 10 0.373 a 0.130 a 0.468 a 0.112 b 0.557 a 0.105 a 0.469 a 0.048 a 0.187 a(0.019) (0.002) (0.011) (0.002) (0.024) (0.002) (0.018) (0.001) (0.005)

P 3 years 20 0.470 b 0.143 b 0.541 b 0.118 a 0.686 b 0.119 b 0.567 b 0.055 b 0.229 b(0.003) (0.001) (0.004) (0.001) (0.005) (0.001) (0.003) (0.001) (0.001)

T1 1–3 generations 10 0.418 a 0.130 a 0.486 a 0.125 a 0.631 a 0.111 a 0.486 a 0.053 a 0.214 a(0.011) (0.002) (0.008) (0.001) (0.016) (0.001) (0.014) (0.002) (0.005)

T1 2 years 20 0.460b 0.132 b 0.499 a 0.126 a 0.692 b 0.117 b 0.557 b 0.058 b 0.239 b(0.005) (0.002) (0.005) (0.001) (0.006) (0.001) (0.005) (0.001) (0.002)

T2 1–3 generations 5 0.428 a 0.140 a 0.509 a 0.127 a 0.644 a 0.120 a 0.517 a 0.050 a 0.222 a(0.004) (0.009) (0.017) (0.001) (0.004) (0.002) (0.007) (0.001) (0.009)

T2 2 years 18 0.483 b 0.140 a 0.569 b 0.128 a 0.725 b 0.124 b 0.605 b 0.054 b 0.249 b (0.004) (0.01) (0.004) (0.001) (0.005) (0.001) (0.005) (0.001) (0.004)

T3 1–3 generations 10 0.459 a 0.148 a 0.534 a 0.129 a 0.693 a 0.117 a 0.551 a 0.055 a 0.243 a(0.011) (0.003) (0.019) (0.001) (0.025) (0.003) (0.011) (0.002) (0.011)

T3 2 years 21 0.450 a 0.142 a 0.549 a 0.125 b 0.651 b 0.118 a 0.532 a 0.054 a 0.244 a(0.005) (0.002) (0.006) (0.001) (0.007) (0.001) (0.005) (0.001) (0.003)

For each clone mean values of specimens followed by different letter are significantly different (P < 0.05). *See Table 2 for explanation of the characters.

Morphological variation of M. persicae populations 241

Fig. 5. The tree obtained for the individuals of 157 clones of Myzus persicae collected from different host-plants. PT-, peach, away fromtobacco-growing regions; Pep, pepper; C, cabbage; Sb, sugarbeet; T, tobacco; PT+, peach in tobacco-growing regions. The threshold forsplitting a node was based on a 1% improvement of the apparent misclassification rate. The symbol 1413PT- in node 6 denotes thatthere are 141 individual aphids from peach away from tobacco growing region at this node.

Table 5. The Mahalanobis distances between clones of Myzus persicae collected from different host-plants and regions of Greece and fromCaserta, Italy.

Region Host-plant 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

1 Lehonia peach –2 Portaria peach 2.6* –3 Falania peach 1.2* 2.0* –4 Velestino pepper 2.0* 4.2* 1.5* –5 Dimini pepper 2.4* 4.7* 2.8* 1.0* –6 Heraklion pepper 4.8* 6.7* 2.5* 1.5* 4.2* –7 Alexandria** cabbage 2.0* 2.0* 1.3* 2.0* 2.2* 4.8* –8 Alexandria** sugarbeet 3.9* 5.4* 5.2* 3.6* 3.0* 8.3* 2.2* –9 Agrinio** tobacco 5.8* 7.6* 3.9* 5.3* 7.2* 6.2* 3.3* 7.7* –10 Alexandria** tobacco 7.0* 10.4* 6.4* 6.7* 7.2* 9.5* 4.2* 7.5* 1.2* –11 Anavra** tobacco 6.8* 9.3 5.3* 5.6* 6.4* 7.3* 3.4* 6.3* 0.8* 0.6* –12 Caserta** tobacco 12.9* 17.2 11.3* 13.6* 16.3* 13.1* 12.9* 20.0* 4.1* 5.4* 6.7* –13 Katerini** tobacco 5.5* 9.5 4.5* 4.5* 4.7* 6.6* 3.7* 7.1* 15* 0.9* 1.1* 6.0* –14 Kilkis** tobacco 5.9* 7.9 4.3* 5.3* 5.6* 7.6* 2.5* 5.7* 1.0* 1.0* 0.5* 7.7* 0.9* –15 Preveza** tobacco 8.8* 13.6 8.6* 9.9* 10.8* 11.3* 7.8* 12.5* 2.3 1.5 2.5 2.7 2.4 3.4* –16 Amphiklia** tobacco 6.7* 9.19 6.7* 6.8* 7.6* 9.6* 3.7* 5.4* 1.7* 1.0* 1.1* 7.6* 2.8* 1.9* 2.7 –17 Tsaritsani** tobacco 6.5* 8.4 4.5* 5.1* 5.3* 7.2* 2.9* 6.3* 1.4* 1.4* 0.7* 8.3* 0.8* 0.1 4.1* 2.8* –18 Alexandria** peach 4.9* 9.0 4.3* 3.9* 5.4* 5.5* 4.3* 5.7* 2.0* 3.1* 2.6* 6.3 2.1* 2.6* 4.7* 3.5* 2.8* –19 Aridea** peach 5.3* 9.0 3.9* 5.5* 6.3* 7.4* 4.1* 9.3* 3.7* 3.6* 3.9* 10.3* 2.2* 2.7* 5.9* 6.3* 2.5* 5.7* –

* Denotes significant difference (P < 0.05). **Denotes tobacco-growing region.

Table 6. The Mahalanobis distances between clones of Myzus persicae collected from different host-plants and regions of Greece and fromCaserta, Italy.

Region Agrinio Amphiklia Anavra Tsaritsani Katerini Alexandria Kilkis

Agrinio –Amphiklia 1.79* –Anavra 0.80* 1.04* –Tsaritsani 1.46* 3.05* 0.81* –Katerini 1.60* 2.71* 1.02* 0.92* –Alexandria 1.36* 0.95* 0.68* 1.69* 0.81* –Kilkis 1.00* 1.97* 0.41* 0.18 0.96* 1.15* –

* Denotes significant difference (P < 0.05).

morphometric studies is whether the variation in charactersreflects the effect of the environment, or is caused by geneticdifferences. In our study all the clones were reared in acommon environment. The clones from tobacco and otherplants maintained their discrete morphology after one tothree years rearing in the laboratory. In addition, clones ofthe tobacco form were collected from peach, where thesexual phase occurs. Therefore, the differences inmorphology between clones feeding on tobacco and onother hosts must be due to genetic differences. Thesefindings agree with those of Blackman (1987), who foundthat a clone of the tobacco form was just as distinct afterrearing for up to seven years on cabbage, and concluded thatclones of the tobacco form share a complex morphometrictrait that is well stabilized and under polygenic control.

In the present study, specimens of both forms reared oncabbage for one generation were smaller than those rearedon other plants for the same period, suggesting immediate

host effects such as nutrition. However, this did not affectscores on CV1, which still separated clones of the twogroups irrespective of host plant on which they were reared.

Another point of discussion is the higher values of CV1exhibited by specimens collected from peach and measuredafter only a short period of rearing in the laboratory,compared with the same clones reared for many generationson secondary hosts. The morphology of the fundatrix, orstem mother, that hatches from an egg on peach, differsgreatly from that of later parthenogenetic generations(fundatrigeniae), and the generation(s) that follow thefundatrix on peach prior to migration to secondary hosthave intermediate morphology. Samples transferred frompeach without a migratory alate generation to the secondaryhost, and reared for only one to three generations on potato,may have retained some degree of ‘fundatriciform’morphology. This indeed would explain the generaldisplacement of CV1 values of peach-collected clones

242 J.T. Margaritopoulos et al.

Fig. 6. Dendrograms of cluster analysis based on Mahalanobis distances between clones of Myzus persicae collected (a) from differenthost-plants and regions of Greece and from Caserta in Italy and (b) from tobacco from different regions of Greece. PT-, peach, away fromtobacco-growing regions; Pep, pepper; C, cabbage; Sb, sugarbeet; T, tobacco; PT+, peach near tobacco-growing regions

compared with those collected on secondary host plants. Achange in morphological characters depending on thenumber of generations after the egg hatch is also found inalate morphs of M. persicae (Woodford, 1977). Despite theseeffects on phenotype, the CV1 values of the tobacco formstill separated it from M. persicae colonizing other secondaryhosts. Examination of the morphology of apterae bymultivariate analysis for successive generations after egghatch would clarify the duration of the ‘fundatrix effect’ andmake it possible to estimate the longer-term effects ofparthenogenesis on phenotype.

The tobacco form may have acquired a differentmorphology, possibly as an adaptation to the morphologicalcharacteristics of the host-plant, and seems able to maintainthis even when at least part of the population has a bisexualphase on peach, indicating that there must be degree ofreproductive isolation. However, this is clearly not complete;for example, DNA sequences of fragments of the E4 and FE4genes, involved in esterase-based mechanisms of insecticideresistance, are identical in both forms, which must be aconsequence of interbreeding (Field et. al., 1994). A low levelof hybridization could explain the intermediate CV1 valuesof some clones from tobacco and other host-plants, andmight also explain the absence of consistent differences inRAPD banding patterns.

Concerning the geographical variation in morphology ofaphids from tobacco, cluster analysis revealed some differ-entiation in phenotype between clones, notably separatingthose collected in the south-east at Amphiklia from those inother regions. The bisexual phase on peach is rare in thesouth, and populations may therefore consist of rather fewgenotypes. The genotypes that were sampled in Amphikliaprobably were under-represented or only poorlyrepresented elsewhere. The nature and degree of the geneticdifferences in population studies is probably underestimatedby examining only the phenotype. Therefore, a molecularapproach (RAPD, mitochondrial or microsatellite DNAanalysis) is needed to clarify the differences in the geneticstructure between the former populations. A previous studywith RAPDs on aphids from tobacco in Greece failed todetect any diagnostic band or an important variation inbanding pattern between clones collected from differentregions (Margaritopoulos et al., 1998). However, thenumbers of clones from each region were not enough to

provide an estimation of genotype diversity within andbetween populations.

A more intensive sampling programme and geneticanalysis of M. persicae populations in Greece is currently inprogress using microsatellite DNA markers.

Multivariate analysis of morphological characters is aneffective tool in solving problems in aphid taxonomy(Blackman et al., 1977; Blackman, 1987; Blackman & Eastop,1987; Lazzari & Voegtlin, 1993). However, the applicationhas limitations concerning the discrimination at the level ofthe individual aphid. Therefore, in systematic or populationstudies it is useful for morphometrics to be used inconjunction with studies of DNA, allozyme electromorphs,geographical isolation, life-cycle characteristics and hostplant preferences.

Acknowledgements

We thank Demetrios Boutos, Nikolaos Kavalieratos andElizabeth Chatzivassiliou for assistance with collection ofaphids. The present study was partly supported by the EPETII 453 grant from the General Secretariat for Research andTechnology of Greece and the Tobacco 96/T/18 from theEuropean Union.

References

Blackman, R.L. (1971) Variation in the photoperiodic responsewithin natural populations of Myzus persicae (Sulzer).Bulletin of Entomological Research 60, 533–546.

Blackman, R.L. (1987) Morphological discrimination of atobacco-feeding form from Myzus persicae (Sulzer)(Hemiptera: Aphididae), and a key to New World Myzus(Nectarosiphon) species. Bulletin of Entomological Research 77,713–730.

Blackman, R.L. (1992) The use of ordination techniques todiscriminate within pest aphid species complexes.pp. 261–275 in Sorensen J.T. & Footit R. (Eds). Ordination inthe study of morphology, evolution and systematic of insects.Amsterdam, Elsevier Science Publishers.

Blackman, R.L. &. Eastop, V.F. (1984) Aphids of the World‘s crops:an identification and information guide. London, John Wiley &Sons Publications.

Blackman, R.L. & Eastop, V.F. (1987) The strawberry aphid

Morphological variation of M. persicae populations 243

Table 7. Fisher‘s linear discriminant functions (LDF) and percentage correct classification for apterous virginoparae of Myzus persicaeusing all nine characters.

Character* Non-tobacco Tobacco form Actual group Predicted group Percentage (%)form correct classified

Non-tobacco TobaccoCoefficient form form

ant III –7.872920 –33.64220 Non-tobacco form 470 98 82.7base VI 583.0210 608.5092 Tobacco form 96 1059 91.7pt 98.15707 120.6124 Total 573 1150 88.7urs 5379.970 5833.843hf –250.0200 –247.8500ht2 1046.127 897.4741Ls 1.925079 8.682135mws 1439.614 1318.625lc 10.04079 25.08850Constant –396.8980 –439.1960

* See Table 2 for explanation of the characters.

complex, Chaetosiphon (Pentatrichopus) spp. (Hemiptera:Aphididae) taxonomic significance of variation inkaryotype, chaetotaxy and morphology. Bulletin ofEntomological Research 77, 201–212.

Blackman, R.L. & Spence J.M. (1992) Electrophoretic distinctionbetween the peach–potato aphid, Myzus persicae and thetobacco aphid, Myzus nicotianae (Homoptera: Aphididae).Bulletin of Entomological Research 82, 161–165.

Blackman, R.L. & Spence, J.M. (1994) The effects of temperatureon aphid morphology, using a multivariate approach.European Journal of Entomology 91, 7–22.

Blackman, R.L., Eastop, V.F. &. Hills, M. (1977) Morphologicaland cytological separation of Amphorophora Buckton(Homoptera: Aphididae) feeding on European raspberryand blackberry (Rubus spp.). Bulletin of EntomologicalResearch 66, 285–296.

Dixon, A.F.G. (1998) Aphid ecology. 2nd edn. London, Chapman& Hall.

Field, L.M., Javed, N., Stribley M.F. & Devonshire, A.L. (1994)The peach–potato aphid Myzus persicae and the tobaccoaphid Myzus nicotianae have the same esterase-basedmechanisms of insecticide resistance. Insect MolecularBiology 3, 143–148.

Gillham, M.C. & Claridge, M.F. (1994) A multivariate approachto host plant associated morphological variation in thepolyphagous leafhopper Alnetoidia alneti (Dahlbom).Biological Journal of Linnean Society 53, 127–151.

Ilharco, F.A. & van Harten, A. (1987) Systematics. pp. 51–77 inMinks, A.K. & Harrewijn, P. (Eds) Aphids. Their biology,natural enemies and control. Volume A, 450 pp. Amsterdam,Elsevier.

Kolesova, D.A., Kuznetova, V.G. & Shaposhnikov, G.K. (1980)Clonal variability in peach aphid, Myzus persicae Sulz.(Homoptera: Aphididae) [in Russian]. EntomologicheskoeObozrenie 59, 514–528. [English translation in Ent. Rev.,Wash. 59, 21–34.].

Krzanowski, W.J. (1990) Principles of multivariate analysis.Oxford, Clarendon Press.

Lazzari, S.M.N. & Voegtlin, D.J. (1993) Morphological variationin Rhopalosiphum padi and R. insertum (Homoptera:

Aphididae) related to host plant and temperature. Annals ofthe Entomological Society of America 86, 26–36.

Margaritopoulos, J.T., Mamuris, Z. & Tsitsipis J.A. (1998)Attempted discrimination of Myzus persicae and Myzusnicotianae (Homoptera: Aphididae) by random amplifiedpolymorphic DNA polymerase chain reaction technique.Annals of the Entomological Society of America 91, 602–607.

Moran, A.N. (1986) Morphological adaptation to host plants inUroleucon (Homoptera: Aphididae). Evolution 40,1044–1050.

Moran, A.N. (1988) The evolution of host-plant alternation inaphids: evidence for specialisation as dead end. AmericanNaturalist 132, 681–706.

Sneath, P.H.A. & Sokal, R.R. (1973) Numerical taxonomy. SanFrancisco, W.H. Freeman & Co.

Takada, H. (1986) Genotype composition and insecticideresistance of Japanese population of Myzus persicae (Sulzer)(Homoptera: Aphididae). Zeitschrift für AngewandteEntomologie 100, 451–458.

Wool, D. (1977) Genetic and environmental components ofmorphological variation in gall-forming aphids(Homoptera: Aphididae, Fordinae) in relation to climate.Journal of Animal Ecology 46, 875–889.

Woodford, J.A.T. (1977) The effect of genotype, environment,age and morph on morphological variation in alate Myzuspersicae (Sulzer) (Hemiptera: Aphididae). Bulletin ofEntomological Research 67, 685–693.

Zagorovskii, A.V. (1947) Leaf peach aphid (Myzodes persicaeSulzer) as a tobacco pest in Uzbekistan and measures for itscontrol [in Russian]. 45 pp. Frunze Zos, VsesoyuznyInstitute Tabaka and Makhorki [cited in Kolesova et al.,1980].

Zintzaras, E., Brown, N.P. & Kowald A. (1994) Growing aclassification tree using the apparent misclassification rate.Computer Applications in the Biosciences 10, 263–271.

(Accepted 14 March 2000)© CAB International, 2000

244 J.T. Margaritopoulos et al.

Bulletin of Entomological Research (2000) 90, 451 451

Corrigendum

Bulletin of Entomological Research, Vol. 90(3), pp. 233–244

The following table replaces that which originally appeared in the above article.

Table 3. Mean lengths (±SE) in mm of nine characters of clones of Myzus persicae reared on different host-plants.

Clones Host-plants ant III* base IV pt urs hf ht2 ls mws lc

T1 potato (n=20) 0.460 b 0.132 ab 0.499 ab 0.126 a 0.692 ab 0.117 b 0.557 b 0.057 b 0.239 b (0.005) (0.002) (0.005) (0.001) (0.006) (0.0006) (0.005) (0.001) (0.002)

cabbage (n=20) 0.423 a 0.123 a 0.486 a 0.127 a 0.613 a 0.113 a 0.407 a 0.054 a (0.208 a (0.007) (0.001) (0.004) (0.001) (0.009) (0.001) (0.008) 0.001) (0.004)

broad bean (n=20) 0.476 c 0.135 b 0.518 c 0.126 a 0.715 b 0.120 c 0.559 b 0.053 a 0.239 b(0.005) (0.001) (0.004) (0.001) (0.005) (0.0006) (0.004) (0.0006) (0.002)

tobacco (n=20) 0.479 c 0.135 b 0.513 bc 0.130 a 0.732 b 0.120 c 0.582 c 0.055 a 0.248 b (0.004) (0.001) (0.006) (0.001) (0.006) (0.0006) (0.004) (0.001) (0.003)

T2 potato (n=18) 0.482 b 0.134 a 0.569 b 0.128 a 0.727 c 0.129 c 0.605 b 0.054 b 0.249 b(0.004) (0.001) (0.004) (0.001) (0.005) (0.001) (0.005) (0.001) (0.003)

cabbage (n=20) 0.382 a 0.131 b 0.512 a 0.127 a 0.573 b 0.112 a 0.456 a 0.052 a 0.203 a(0.005) (0.0007) (0.005) (0.001) (0.006) (0.001) (0.006) (0.001) (0.003)

broad bean (n=20) 0.483 b 0.147 c 0.521 a 0.127 a 0.694 a 0.118 b 0.562 c 0.052 ab 0.246 b (0.005) (0.001) (0.008) (0.0005) (0.005) (0.001) (0.005) (0.0006) (0.003)

tobacco (n=20) 0.475 b 0.143 d 0.550 b 0.127 a 0.738 c 0.123 c 0.609 b 0.054 b 0.270 c (0.006) (0.002) (0.009) (0.001) (0.009) (0.001) (0.005) (0.0005) (0.003)

T3 potato (n=21) 0.450 c 0.142 a 0.549 b 0.125 a 0.651 d 0.118 c 0.532 d 0.054 b 0.244 c (0.005) (0.002) (0.006) (0.001) (0.007) (0.001) (0.005) (0.001) (0.003)

cabbage (n=22) 0.369 a 0.138 a 0.515 a 0.125 a 0.565 a 0.111 a 0.436 a 0.050 a 0.206 a (0.006) (0.001) (0.004) (0.0005) (0.007) (0.001) (0.007) (0.001) (0.003)

broad bean (n=19) 0.417 b 0.139 a 0.513 a 0.122 b 0.587 b 0.109 ab 0.474 c 0.050 a 0.208 a (0.007) (0.002) (0.006) (0.001) (0.009) (0.001) (0.006) (0.001) (0.004)

tobacco (n=20) 0.424 b 0.139 a 0.526 a 0.125 a 0.626 c 0.113 b 0.497 b 0.052 ab 0.230 b (0.003) (0.002) (0.009) (0.0005) (0.005) (0.0005) (0.009) (0.001) (0.002)

P potato (n=20) 0.470 b 0.143 b 0.541 c 0.119 b 0.686 b 0.119 b 0.567 b 0.055 b 0.229 b (0.003) (0.004) (0.004) (0.0006) (0.005) (0.001) (0.003) (0.0008) (0.001)

cabbage (n=11) 0.376 a 0.134 a 0.488 a 0.117 a 0.544 a 0.111 a 0.424 a 0.051 a 0.179 a (0.007) (0.005) (0.008) (0.0006) (0.007) (0.001) (0.006) (0.001) (0.003)

broad bean (n=22) 0.476 b 0.141 b 0.528 b 0.119 b 0.680 b 0.117 b 0.554 b 0.055 b 0.224 b (0.007) (0.006) (0.004) (0.0006) (0.01) (0.001) (0.009) (0.0006) (0.003)

T1 and T2: red androcyclic clones collected from tobacco from Anavra region. T3: green holocyclic clone collected from tobacco fromAlexandria region. P: green holocyclic clone collected from peach, away from tobacco growing regions (Portaria, Volos). For each clone mean values of specimens followed by different letter are significantly different (P<0.05). Duncan’s test was used *See Table 2 for explanation of the characters.