One aphid species induces three gall types on a single plant: Comparative histology of one genotype and multiple extended phenotypes

Flora 210 (2015) 19–30

Contents lists available at ScienceDirect

Flora journal homepage: www.elsevier.com/locate/flora

One aphid species induces three gall types on a single plant: Comparative histology of one genotype and multiple extended phenotypes L. Kurzfeld-Zexer a , S. Lev-Yadun b , M. Inbar a,∗ a b

Department of Evolutionary & Environmental Biology, University of Haifa, Mount Carmel, Haifa 3498838, Israel Department of Biology & Environment, Faculty of Natural Sciences, University of Haifa-Oranim, Tivon 36006, Israel

a r t i c l e

i n f o

Article history: Received 6 March 2014 Received in revised form 28 September 2014 Accepted 7 October 2014 Edited by Shahin Zarre Available online 1 November 2014 Keywords: Gall-forming aphids Phenotypic plasticity Pistacia atlantica Smynthrodes betae Sclereids

a b s t r a c t In aphids, usually only the fundatrix (F1) induces galls. In Smynthrodes betae (Fordini) however, three gall types may coexist on the leaflets of a single host species, Pistacia atlantica (Anacardiaceae). The fundatrix induces pea-shaped galls on the leaflet midvein early in spring and its genetically identical daughters (F2) disperse from these galls and induce spindle-shaped galls on leaflet margins. In some of the fundatrix’ galls, the daughters (F2) continue to reproduce and complete their normal life cycle within the same gall (CLC galls). Using a comparative anatomy approach, the aim of this study is to evaluate the role of different aphid generations and gall location in controlling gall structure. Histological sections of the three gall types as well as galled and ungalled leaflets were examined. The leaflet’s mesophyll is highly modified in all gall types expressing hyperplasia and hypertrophy and collateral veins occupy the inner part of the gall’s wall. The epidermis lining the chamber of only the F1 and the CLC galls is covered with trichomes. The F1 galls have a rather simple structure, composed mostly of large parenchyma cells. The F2 and CLC galls have two parenchyma layers separated by a sclereid layer that creates a protective hardened structure. The structural similarity of the F2 and the CLC galls indicates that most histological modifications are controlled by the aphids and depend on either the continuous or cumulative activity of the following aphid generations in the gall. © 2014 Elsevier GmbH. All rights reserved.

Introduction Galls are abnormal plant structures induced by various organisms, in particular insects (Mani, 1964). The mechanism of gall induction by insects remains unknown, but it seems that the insect controls gall development, subject to suitability and reactivity of the plant tissues (Weis et al., 1988). This hypothesis is supported by phylogenetic studies of several lineages of gall-forming insects (Cook and Gullan, 2008; Inbar et al., 2004; Nyman et al., 2000; Stern, 1995; Stone and Cook, 1998). Galls are therefore described as the extended phenotypes of the insects (Dawkins, 1982). The galls serve as an “incubator” for the insects that may gain better nourishment and protection from unfavourable abiotic conditions and natural enemies (Price et al., 1987). Most insect galls are highly organised structures; defined as histoid prosoplasmic galls (sensu Küster, 1911). They have a definite size and shape, often symmetrical, and usually display a novel tissue

∗ Corresponding author. Tel.: +972 4 828 8767. E-mail address: [email protected] (M. Inbar). http://dx.doi.org/10.1016/j.flora.2014.10.007 0367-2530/© 2014 Elsevier GmbH. All rights reserved.

pattern compared to ungalled organs (Chakrabarti, 2007; Raman, 2007). However, the complexity of gall structure varies enormously among and within insect groups (Raman, 1996; Rohfritsch, 1992), ranging from local overgrowths (e.g., Rey, 1992) to highly differentiated structures with orderly arrangement of various types of cell layers as exhibited by cecidomyiids and cynipid wasps (Dorchin et al., 2002; Rey, 1992; Rohfritsch, 1992; Sliva and Shorthouse, 2006). A nutritive tissue may develop in the inner layer of some insect induced galls (e.g., Bronner, 1992; Rohfritsch, 1977). A lignified sclerenchyma sheath, which provides physical protection may be formed external to the nutritive tissue (e.g., Dorchin et al., 2002; Rohfritsch, 1992). Usually each insect species induces a single gall type. In some cases two gall types are induced by the same species, on different host plant or organs, or at different periods (e.g., Ananthakrishnan, 1992; Ananthakrishnan and Raman, 1989; Shorthouse and Rohfritsch, 1992). The induction of two gall types extends the ability of the insect to exploit variable plant resources (Dorchin et al., 2009; Miller, 1998; Rhomberg, 1980; Wool and Burstein, 1991). In cynipid wasps, distinct generations of the same species may induce different galls on different plant organs (e.g.,

20

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

Rey, 1992). Different plant organs (galling sites) might restrict gall development (Wool, 1997) and thus promote gall dimorphism. Dimorphic galls can also be related to different induction signals given by the insects. For example, some scale insects (Coccidea) induce sexually-dimorphic galls, often on the same plant organ, suggesting that different signals are involved (Gullan et al., 2005). The longer feeding period and the sessile habit of the coccoid female may also influence gall size and structure (Gonc¸alves et al., 2005). The development of sexually dimorphic gall structures in the pteromalid wasps was related to cytokinins and auxins levels (Dorchin et al., 2009). Aphids (Hemiptera: Aphididae) induce a variety of galls with different morphologies and structural complexities (Inbar et al., 2004). Aphids in the tribe Fordini induce galls on Pistacia spp. (Anacardiaceae) hosts (Wool, 1995). Their life cycle is complex and includes alternation between a primary host (Pistacia spp.) and roots of secondary hosts, on which they do not induce galls. Aphids are phloem-feeders and their galls may have an extensive internal vascularisation (Wool et al., 1999). In most gall-forming aphids, only the fundatrix (F1) is capable of inducing a gall despite the genetic identity with her offspring (Moran, 1988; Wool, 2004). This paper deals with one of the most common Fordini species, Smynthurodes betae Westw., which forms galls on leaflets of P. atlantica Desf. The fundatrix induces small, pea-shaped (“temporary”) galls on leaflet midveins early in spring (Fig. 1A). About three weeks later, her offspring (F2) induce different (“final”) galls on leaflet margins of adjacent young leaves (Fig. 1B). Usually, the fundatrix dies within a few weeks but the empty temporary galls may remain on the tree until leaf-fall in autumn (Wertheim, 1954). Each final gall is induced by a single F2 aphid in which two more parthenogenetic generations are produced. In autumn, winged migrants (F4) disperse from the F2 gall to various secondary hosts (see Fig. 1 in Wool and Burstein, 1991). The F2 galls contain about 30 winged aphids. In S. betae there is a further complication, since in some of the fundatrix galls, the daughters (F2) continue to reproduce and complete their life cycle within this temporary gall. The “CLC” (complete life cycle) galls contain about twelve aphids and they are quite rare, comprising between 3 and 24% of the fundatrix galls (Wool and Burstein, 1991). Morphologically they are not easily distinguished from F1 galls. Usually they appear to be darker and larger (Fig. 1C) and become harder than normal fundatrix galls during the summer (Inbar personal observations; Wool and Burstein, 1991). This is a rare example of induction of three gall types by the same species on a given host plant (see also Gonc¸alves et al., 2009). Pistacia atlantica is a dioecious and deciduous tree with an IranoTuranian distribution (Zohary, 1952). Inter-population variability in some leaf morpho-anatomical characters has been reported (Ait Said et al., 2011; Belhadj et al., 2007; Inbar and Kark, 2007). For example, the thickness of the epidermis, cuticle, palisade parenchyma and total lamina as well as the density of epidermal trichomes increased with the degree of aridity (Ait Said et al., 2011). Histological studies in the Fordini are limited to galls induced on P. palaestina and P. terebinthus (Álvarez et al., 2009; Álvarez, 2011a, 2011b; Wool and Bar-El, 1995; Wool et al., 1999). These galls induced on the same plant organ and even at the same galling site (e.g., leaflet margins) exhibit different histological modification of the ungalled leaflet tissues (e.g., Álvarez et al., 2009). Here we present a comparative histological study of three gall types induced by a single aphid species (S. betae). Each gall is induced and/or inhabited by different generations on the leaflets (leaflet midvein vs. leaflet margin). This system provides an opportunity to evaluate the role of different aphid generations and location in the modifications that occur in the plant tissues during gall formation.

Material and methods Plant material Galls induced on leaflets of P. atlantica by the aphid S. betae and ungalled control leaves were collected from trees growing in several sites in northern Israel: Kiryat Tiv’on; Beit Lehem HaGlilit; Gamla and Bar’am. Seven fundatrix (hereafter, F1) and 11 final (hereafter, F2) galls were sampled during August-September from four and five P. atlantica trees, respectively. Six CLC galls were collected in late June from two trees. All galls were collected with the galled leaf (control 1). In addition, another ungalled leaf from the same branch was sampled as a second control (control 2). Samples were kept cool from the time of collection until their arrival at the laboratory (for a few hours at the most) and then fixed in 96% ethanol and glacial acetic acid (ratio 3:1). All fixed samples were kept in the fixative solution for at least several months. Histology The histological procedures were performed according to Ruzin (1999) with few modifications. The fixed samples were cut transversely with a sharp razor blade into smaller segments (∼1 cm long). Three types of samples were prepared: galls only, galls with their attached galled leaflets and ungalled control leaflets. These samples were placed in 20 ml vials with 50% ethanol overnight and then dehydrated in a series of ethanol and tert-butanol solutions (% ethanol/% tert-butanol: 40/10; 50/20; 50/35; 45/55; 25/75; 0/100; 0/100; 0/100) for at least 8 h in each stage. After dehydration, the samples were infiltrated with paraffin for over one week in an oven at 60 ◦ C. The paraffin was replaced every two days and then the samples were embedded in paraffin until complete solidification. These paraffin blocks were used to prepare 8–10 ␮m thick serial sections of the whole gall (data not shown) on a rotary microtome (Leica RM 2135). Histological sections were stained with Safranin and Fast-Green and mounted permanently on microscope slides with Entellan New (MERCK). Microscopy and measurements The sections were examined under an Olympus BX61 microscope equipped with an Olympus DP70 digital camera under bright-field, fluorescence and polarised light illumination and also under a bright-field Olympus CX21 microscope. Measurements ˆ software. were done using the AnalySISB The following parts were studied (Fig. 1D–F): (1) The external epidermis of the gall, derived from the abaxial leaflet epidermis; (2) the internal epidermis of the gall, derived from the adaxial leaflet epidermis; (3) the modified mesophyll between both epidermal layers; (4) the zone where the gall’s lumen is closed and a narrow opening (ostiole) is formed. Henceforth, the term “ostiolar zone” refers to the gall’s walls adjacent to the ostiole; (5) the midvein; (6) the part of the leaflet lamina adjacent to the ostiolar zone, which separate between the galled tissue and a normal leaflet tissue (transition zone). In addition, we examined different parts of the galled leaflets, located at increasing distance from the transition zone (galled leaflet 1, 2, 3, Fig. 1A–C). Not all zones were found in all sections. The thickness (in ␮m) of the gall’s wall and the leaflet lamina were measured in the different zones (i.e., the main part of the gall (excluding the ostiolar zone), ostiolar zone, transition zone, galled ˆ software. The numleaflet, ungalled leaflet) using the AnalySISB ber of cell layers that composed each zone was counted and the quotient between the thickness (in ␮m) and the number of cell layers was used to evaluate cell size (height). Similar measurements

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

21

Fig. 1. (A–C) Macroscopic views of S. betae galls and galled leaflets (control 1). (A) F1. (B) F2. (C) CLC. The numbers and lines mark the parts of the galled leaflets that were studied. (D–F) Cross sections of the galls. The parts studied in the present work are indicated. (D) F1. (E) F2. (F) CLC. (G–K) Cross sections of ungalled (control 2) P. atlantica leaflet showing different zones. (G) Part of a leaflet lamina including the midvein and medium sized veins. (H) Detailed section of the lamina zone, close to the leaflet centre, showing adaxial and abaxial epidermises and the mesophyll between them. The mesophyll is divided into one layer of palisade parenchyma on the upper side and four layers of spongy parenchyma. (I) The distal margin of the leaflet, showing a decrease in the number of spongy parenchyma layers. (J) Section of the midvein zone showing a group of five vascular bundles, one of them (the upper) is in an inverted position to the others. Arrowheads mark the xylem. The resin ducts are located within the phloem. (K) The lamina close to the midvein zone, under fluorescence light. The xylem, the phloem fibres and the cuticle, especially in the midvein zone, are illuminated. Abbreviations: Abaxial epidermis (epB), Adaxial epidermis (epD), crystal (cr), cuticle (cu), external epidermis (epE), fibres (fi), galled leaflet (GL), internal epidermis (epI), lamina (LA), lumen (L), medium sized vein (mdv), midvein (MIV), modified mesophyll (Mm), ostiolar zone (OZ), palisade parenchyma (pp), resin duct (rd), small vein (sv), spongy parenchyma (sp), transition zone (TZ), xylem (xy). Scale bars: (A–C) = 1 cm; (D–G) = 1 mm; (H–K) = 200 ␮m.

were taken in specific tissues of the galls and leaflets (spongy and palisade parenchyma, sclerid layer and cuticle). Each measurement was taken in 2–5 places, according to the length of the measured zone, and the mean value was used for further statistics. In addition, measurements of the diameter (in ␮m) of the five largest vessel members and of all (3–5) resin ducts were taken from the midvein of each gall type, galled leaflets and ungalled leaflets.

Statistical analyses Differences between gall types, galled and ungalled leaflets, and between different zones of the same gall were tested using one-way ANOVA followed by Post-Hoc (Tukey) tests. A two-way ANOVA was used to evaluate the effect of gall type and the zone on the relative thickness of the sclereid layer. All data were tested for normality prior to statistical analyses using the Kolmogorov–Smirnov test

22

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

Table 1 Characteristics of different gall types induced by S. betae as compared with ungalled P. atlantica leaflet. (+) Presence; (−) absence. The “relative amount” is a subjective observation of the investigator, where “1” is the lowest presence and “3” the greatest.

General characteristic Location Shape Unsealed gall Toughness External/Abaxial epidermis Type Form of the cells Cuticle Trichomes Internal/Adaxial epidermis Type Form of the cells Cuticle Trichomes Modified/non galled mesophyll Form of parenchyma cells Amyloplasts (relative amount) Crystals (relative amount) Sclereids Vascular bundles Collateral bundles Xylem oriented towards Resin duct (in medium-sized veins) Midvein Part of the gall Location in galls Form of xylem Phloem fibres encircle the vascular tissue Intraphloem resin duct Sclereids Ostiolar zone Flattening of the zone Sclereids Trichomes Transition zone Division to palisade & spongy parenchyma Small cells parenchyma

F1

F2

CLC

Leaflet midvein Pea-shaped + −

Leaflet margin Spindle-shaped + +

Leaflet midvein Pea-shaped + +

Uniseriate Cuboidal/squamous Thin Scarce

Uniseriate Cuboidal/squamous Conspicuous Few

Uniseriate Cuboidal/squamous Thin Scarce

Uniseriate Cuboidal Very thin or absent Scarce

Uniseriate Cuboidal/irregular Absent Many

Uniseriate Squamous Thin or absent None

Uniseriate Cuboidal/irregular Absent Many

Uniseriate Cuboidal Very thin or absent Few

Polyhedral 1 1 −

Polyhedral 2 1 +

Polyhedral 1 1 +

Long and narrow − 2 −

+ Lumen +

+ Lumen +

+ Lumen +

+ Adaxial side +

+ Centre Whole ring/arched − + -

+/− Ostiolar zone Proximal side Sectorial +/− + +

+ Centre Whole ring/arched − + +/−

Sectorial + + −

+ − +

+ + −

+ + +

− +

− +

− +

for normality. Since only two trees were sampled for CLC galls, we ignored the “tree” factor and all data for each gall type were pooled for the statistics analysis. The sample sizes were variable and they are given separately for each case (in the Results section). All tests were conducted with the PASW SPSS statistics 17 software. Results Anatomy of ungalled leaflets Ungalled P. atlantica leaflets (control 2) have a dorsiventral leaf structure (Fig. 1G–I). The epidermis is uniseriate on adaxial and abaxial sides. The abaxial epidermis cells seem to be smaller than the adaxial cells and they are both usually cubic in shape with a thin cuticle (Fig. 1H and I). Only a few (usually none) trichomes were observed. Internal to the epidermis is the photosynthetic mesophyll, which is composed of one layer of dense palisade parenchyma on the adaxial side and 3–4 layers of spongy parenchyma on the abaxial side (Fig. 1H). The palisade parenchyma accounted on average for 38% of the total lamina thickness and about 45% of the mesophyll thickness. The cells in the spongy parenchyma are long and narrow though shorter than the palisade cells. Large intercellular spaces occur mostly in the spongy parenchyma (Fig. 1H). Crystals of calcium oxalate, usually in the shape of druses, are well noticeable, especially in the spongy parenchyma close to the vascular bundles (Fig. 1H). The vascular system of the leaflet includes the midvein, medium-sized and minor veins distributed along the leaflet lamina

Leaflet

(Fig. 1G, J and K). The bundles are collateral. The xylem and phloem are oriented towards the adaxial and abaxial sides, respectively. The midvein’s vascular tissue is composed of four to five vascular bundles. One of these bundles is located in an adaxial position relative to the others (usually the smallest) and it was always inverted (i.e., the xylem is oriented towards the abaxial side) (Fig. 1J). The midvein vascular tissue is encircled by fibres (Fig. 1K). Resin (schizogenous) ducts are located within the phloem in the midvein and the medium-sized veins (Fig. 1G and J) but not in the minor veins.

Anatomy of S. betae galls All gall types are hollow and unsealed (Fig. 1D–F) expressing vast modifications of the leaflet tissues (summarised in Table 1). The F1 and the CLC midvein galls comprise a folding of both margins of the adaxial surface of the leaflet towards each other and an invagination of the midvein zone towards the abaxial side. Thus, these galls protrude on the underside of the leaflet, with narrow opening on the upper side (Fig. 1A and C). The F2 galls comprise a folding of the leaflet margin towards the adaxial face of the leaflet. The gall includes the midvein in its proximal side, which is closer to the petiole, while its distal part does not include the midvein (Fig. 1B). Although the origin, external morphology and induction site of CLC galls are similar to F1 galls, their internal wall structure (described below) is more similar to F2 galls than to F1 galls.

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

23

Fig. 2. (A–C) Cross sections in the walls of S. betae different gall types including both epidermal layers and the modified mesophyll between them. Note the similarities in the histologic characters between the F2 (B) and the CLC (C) gall types, which generally differ from the F1 (A) gall type. (D–F) Sections under fluorescence light, emphasizing the presence of trichomes (tr) only in the internal epidermis of the F1 (D) and the CLC gall types (F) and the presence of sclereids (sc) only in the F2 (E) and the CLC (F) gall types. Differences in cuticle thickness of the external epidermis among gall types are noticed (A–F). Reddish-purple inclusions (presumably polyphenols) are seen in the external parenchyma cells as well as in the external epidermis and the sclereids of the F2 (B) and the CLC (C) galls. The appearance of these inclusions in the parenchyma cells of the F1 galls is less uniform (A and G, arrows) than that of the F2 and CLC galls. (H) Parenchyma with amyloplasts in the inner part of the F2 gall type. (I) A close look at the sclereid cells in F2 gall. The pits are easily noticed (arrows). (J and K) Polarisation microscopy showing the sclereids layer in the F2 (J) and in the CLC (K) gall types. (L and M) A non-continuous sclereid layer in CLC galls. Sclereids are not formed in the midvein area (L) or exceptionally are found only near the ostiolar zone (M). Note the peripheral location of the sclereids in section (L). Abbreviations: amyloplasts (am), cuticle (cu), external epidermis (epE), internal epidermis (epI), medium sized vein (mdv), midvein (MIV), lumen (L), ostiolar zone (OZ), paremchyma (pa), resin duct (rd), sclereids (sc), trichomes (tr). Scale bars: (A–H, L, M) = 200 ␮m. (I) = 50 ␮m. (J and K) = 1 mm.

24

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

Table 2 Measurements in the walls of the three gall types of S. betae as compared with the ungalled P. atlantica leaflet. Means ± SE’s of the total thickness (␮m), the number of cell layers, the mean cell height (␮m) and the cuticle thickness (␮m) in the galls’ walls and in the leaflet lamina. Sample sizes for measurements of wall thickness and cell number and height are indicated near each gall type (upper row). Sample sizes for the cuticle measurements are indicated in the lower row. In each line, different letters indicate significant differences.

Total thickness Number of cell layers Mean cell height Cuticle thickness

F1 (n = 7)

F2 (n = 11)

CLC (n = 6)

Ungalled leaflet (n = 9)

579.38 ± 37.524a 15.17 ± 0.753a 38.34 ± 1.927a 7.03 ± 0.322a (n = 5)

418.28 ± 19.921b 17.25 ± 0.559b 24.37 ± 1.047b 14.90 ± 1.282b (n = 11)

497.45 ± 29.806ab 15.30 ± 0.626ab 32.78 ± 1.326bc 10.99 ± 3.224ab (n = 4)

196.48 ± 13.019c 6.22 ± 0.085c 31.50 ± 1.918c Not measurable

Epidermis The external and the internal epidermal layers are uniseriate in all gall types (Fig. 2A–C). The shape of external epidermis cells is similar in all gall types (Table 1). The cells of the internal epidermis are small, squamous and with no trichomes in the F2 galls, while in the F1 and the CLC galls they are larger, have cuboidal or irregular shapes and many trichomes (Fig. 2A–F). The cuticle of the external epidermis is thickest in the F2 galls and thinnest in the F1 galls but always thicker than in the ungalled leaflet (Table 2; one-way ANOVA: F2,19 = 6.06; p = 0.01). The cuticle of the internal epidermis was observed only in the F2 galls, where it was usually thin and appeared as not continuous (Fig. 2B). The modified mesophyll In all three gall types the mesophyll is highly modified and the palisade and spongy parenchyma are indistinguishable. The parenchyma cells are polyhedral in shape, but their size and content differ among galls (Table 1). The F1 galls usually have only large parenchyma cells, which compose most of the gall’s wall. Reddishpurple inclusions appear mostly in the external parenchyma cells (closer to the external epidermis) but do not form a continuous layer along the gall’s wall. Their abundance and pattern of distribution vary among samples (Fig. 2A and G). The F2 and the CLC galls are characterised by two layers (outer and inner) of smaller parenchyma cells, which are separated by a sclereid (“stone cell”) layer (see below). The external parenchyma layer (and some of the sclereid cells) usually contains dense reddish-purple inclusions (presumably polyphenols), which are scarce in the internal parenchyma (closer to the lumen) (Fig. 2B and C). Amyloplasts (light bluish grains inside the cells, Fig. 2H) are observed in the internal parenchyma layer of F2 galls, are less common in the F1 and the CLC galls and are not observed in the control leaflets. In all gall types, small and medium-sized collateral veins are formed in the internal part of the gall’s wall among the parenchyma cells, with their xylem oriented towards the gall lumen (Fig. 1D–F). Intraphloem resin ducts are found in the medium-sized veins (Fig. 2B–C). Sclereids, which provide hardness, develop in F2 and CLC galls (Fig. 2B, C, E, F, H–M), but never appeared in intact leaflets or F1 galls. About 30% of the cell layers in the gall’s wall of both the F2 and CLC gall types are made of sclereids, which compose 35–40% of the total wall thickness. In F2 galls they always comprise a continuous layer, positioned between the two parenchyma layers (Fig. 2B and J), while in CLC galls they may form a non-continuous layer; missing, for example, from the midvein area (Fig. 2L and M). In CLC galls, the sclereids may also comprise the outermost wall layer, just beneath the external epidermis and in such cases the outer layer of the parenchyma is absent (Fig. 2L). Overall, the walls of the three gall types are significantly thicker (one-way ANOVA: F3,32 = 116.17; p  0.01) and have more cells (one-way ANOVA: F3,32 = 92.12; p  0.01) compared with leaflet lamina (Table 2), indicating a process of hyperplasia during gall formation. The mean cell height was significantly higher in F1 galls

and lower in F2 ones, compared with the leaflet lamina (Table 2; one-way ANOVA: F3,32 = 12.1; p  0.01). However, the cell width in all gall types (especially in F1 parenchyma cells), appears larger than in the leaflet lamina (Fig. 2A–C), suggesting that hypertrophy also occurs during gall formation. F1 galls have the thickest wall with the lowest number of cell layers whereas the F2 galls have the thinnest wall with the greatest number of cell layer. Thus, the mean cell height is maximal in F1 and minimal in F2 galls. The CLC galls display an intermediate value of wall thickness although the number of cell layers is similar to that of the F1 galls. Consequently, their mean cell height is also intermediary (Table 2). Ostiolar zone The walls of the three gall types become thinner towards the ostiolar zones (Fig. 3, F1 gall type: one-way ANOVA: F3,18 = 39.7, p < 0.01; F2 gall type: one-way ANOVA: F4,42 = 39.5, p < 0.01; CLC gall type: one-way ANOVA: F3,16 = 32.9, p < 0.01). The walls adjacent to the ostiole of F1 and CLC galls may be non-symmetric, i.e., thinner in one (no specific) side (Fig. 4A, D, H). In F1 galls, the walls in the ostiolar zone are composed of epidermis layers and the parenchyma cells, as in the rest of the gall (Fig. 4A). Sclereids are a major component of the walls adjacent to the ostiole in F2 and CLC galls (Fig. 4B, C, F, G), accounting for about 55% of the number of cell layers and 65% of the wall’s thickness. The internal epidermis cells might be lignified in the ostiolar zone of these galls (Fig. 4C and G). The relative thickness (% of total wall thickness) of the sclereid layer was similar in F2 and the CLC galls. However

Fig. 3. Differences in the total wall or lamina thickness between different zones of the gall, galled leaflet outside the gall region (control 1) and ungalled leaflet (control 2) in the three gall types induced by S. betae. Abbreviations: gall (G), ostiolar zone (OZ), transition zone (TZ), galled leaflet (GL), ungalled leaflet (UGL). Measurements of the gall (G) include all parts of the gall wall, excluding the ostiolar zone (OZ) which was examined separately. Each gall was tested separately. Large letters refer to the F1 galls. Small letters refer to F2 galls and small letters with an apostrophe refer to CLC galls. Different letters indicate significant differences. The numbers in brackets are sample sizes at each measured zone. All measured parts of the galled leaflets were combined for each gall type.

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

25

Fig. 4. (A–H) Cross sections in the ostiolar zone of S. betae gall types: (A and E) F1. (B and F) F2. (C, D, G, H) CLC. Sclereids occupy a major part of the ostiolar zone only in the F2 and the CLC gall types (B–D, F–H). Trichomes appear only in the F1 (E) and in the CLC (C, G) gall types. Differences in wall thickness and sclereids appearance between the two sides of the gall wall are displayed in the F1 (A) and in the CLC (D, H) gall types. (I) Cross section in the proximal side of F2 gall including part of the galled leaflet. The midvein is a part of the ostiolar zone. (J) Cross section of two F2 galls formed on the same leaflet. The midvein is a part of the shared ostiolar zone of both galls. (K) Cross section of the most proximal part of a F2 gall. Sclereids compose a considerable part of the gall tissue and are abundant in the midvein attached to it. (L–Q) Cross sections of the midvein zone of S. betae gall types. (L, O) F1 gall type: The midvein is characterised by a whole secondary xylem ring (arrowheads), surrounded by a phloem ring. Phloem fibres are formed in one area in external (abaxial) position, far away from the gall lumen. (M and P) F2 gall type: The xylem is sectorial (arrowheads), sclereids are abundant and located in different parts of the midvein. (N and Q) CLC gall type: The xylem exhibits almost a whole ring pattern (arrowheads). Phloem fibres are formed as in the F1 galls. Sclereids appear external to the midvein. (A–D, I, L–N) Bright field microscopy; (E, G–H, J–K, O–Q) fluorescence light microscopy; (F) polarisation microscopy. Abbreviations: external epidermis (epE), fibres (fi), gall (G), galled leaflet (GL), lumen (L), midvein (MIV), ostiole (os), paremchyma (pa), phloem (pl), resin duct (rd), sclereids (sc), trichome (tr), xylem (xy). Scale bars: (A–H, K–Q) = 200 ␮m. (I–J) = 1 mm.

26

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

Fig. 5. The relative thickness (% of total wall thickness) of the sclereids layer in the osliolar zone (grey bars) compared with the rest of the gall (white bars) in the F2 and the CLC gall types induced by S. betae. different letters indicate significant differences. F2: n = 11; CLC: n = 5.

in both gall types, sclereids were significantly more abundant in the ostiolar zone than in the rest of the gall’s wall (Fig. 5, two-way ANOVA: gall type: F1,28 = 0.705, NS; zone measured: F1,28 = 44.7, p < 0.001; gall type X zone: F1,28 = 0.272, NS). Trichomes are found in the internal epidermis of the ostiolar zone of F1 and CLC galls (towards the ostiole), but not in F2 galls (Fig. 4E and G). Their density seems to be similar in the ostiolar zones and in other parts of the internal epidermis of these two gall types. Midvein In the F1 and the CLC galls the midvein is always a part of the gall, located in the centre or close to its centre (Fig. 1D and F). The midvein of the F2 galls is included only in the proximal side of the gall (closer to the petiole), being a part of the ostiolar zone (Fig. 4I and J). Unlike the sectorial xylem that is found in the leaflet (Fig. 1J and K) and in F2 galls (Fig. 4M and P), the midvein in F1 and CLC galls is often characterised by a complete or nearly complete xylem ring, surrounded by a phloem ring (Fig. 4L, N, O and Q). The mean diameter of the five largest vessels of the midvein was similar in all gall types, in the part of the galled leaflet outside the gall region and in ungalled leaflet midvein (Fig. 6, one-way ANOVA: F4,28 = 0.308, NS). The phloem fibre ring that encircles the vascular bundles of the leaflet’s midvein (Fig. 1K) is absent in F1 and CLC galls. Instead, the fibres in these gall types are concentrated in one area, always in the external side of midvein, closer to the external epidermis (Fig. 4O and Q). In F2 galls, fibres usually accompany the external bundles,

oriented towards the abaxial side of the midvein, but in most cases they are missing from the inverted bundle oriented towards the ostiole (Fig. 4J and P). Sclereids are abundant in the midvein, as part of the ostiolar zone, in the F2 galls. They occur around or between the vascular bundles of the midvein (Fig. 4M and P). In the gall’s tip, closest to the connection with the petiole, sclereids compose a considerable part of the gall tissue and are abundant in the adjacent midvein (Fig. 4K). In the midvein part that remains outside the gall, sclereids are not found, even at a very short (∼0.5 mm) distance from the gall. In the CLC galls, sclereids might occur in the midvein area, although usually they are found external to the midvein and not as part of the midvein itself (Fig. 4N and Q). Resin ducts appear within the phloem tissue in all gall types, as they do in the intact leaflet. In F1 and CLC galls they are always formed in the external part of the midvein, away from the gall’s lumen (Fig. 4L and N). The resin ducts’ mean diameter was significantly larger in F1 galls as compared with that of the galled leaflet, outside the gall region and that of the ungalled leaflets (Fig. 6, one-way ANOVA: F4,28 = 4.995, p < 0.01). Transition zone The transition zone is characterised primarily by the absence of palisade and spongy parenchyma, which is replaced by small (∼ 20 ␮m in diameter) parenchyma cells (Fig. 7A–G). The length of the transition zone (starting right outward to the gall tissue and ending where a normal leaflet tissue appears) varied among samples; though generally it is rather narrow (no more than 1 mm). In F1 and CLC galls, when only the distal margin of the leaflet remains outside the gall, it comprises the transition zone. When a larger portion of the galled leaflet remains ungalled, a normal leaflet tissue appears distal to the narrow strip of the transition zone (Fig. 7A and C). In F2 galls the transition zone is located right after the midvein (proximal part) or before it (distal part, e.g., Fig. 7H and I). In some cases, the midvein itself (as part of the ostiolar zone) separates between the gall and the intact leaflet tissues (Fig. 4I). In the gall’s distal part, the transition to a leaflet tissue commonly occurs right after a medium-sized vein (Fig. 7G, I–L). In all gall types, the lamina thickness in the transition zone was similar to that of more distant parts of the galled leaflet (Fig. 3). Thus, in all gall types studied the influence of the aphids on the galled leaflet is local. Neither qualitative nor quantitative differences were found between the structure of galled leaflets outside the transition zone and that of ungalled leaflets (Figs. 3 and 6). Discussion Here we describe three histologically-distinct gall types induced by a single aphid species on the leaflets of a given plant species. Despite the similar external morphology of F1 and CLC galls, the latter (as the F2 galls) exhibit a more complex tissue differentiation and organisation (Fig. 2A–C). The influence of the aphids is confined mostly to the gall itself and slightly to a narrow (1 mm at the most) lamina sector adjacent to gall tissue (i.e., the transition zone). Thus, S. betae has a rather local effect on the developmental processes of the host plant (see also Álvarez et al., 2009). Other species that induce much larger galls may have an extended effect on the plant (Aloni et al., 1989; Inbar et al., 2004). Histological modifications and their possible adaptive value

Fig. 6. Measurements in the midvein in different gall types of S. betae, in the part of the galled leaflet outside the gall region and in ungalled leaflets. Means ± SE’s of the diameter of the five largest vessel members and all (3–5) resin ducts. In each group, different letters indicate significant differences. Numbers in columns are sample size.

The epidermis The internal epidermis of F1 and CLC galls is covered with trichomes, which are normally scarce in both epidermal sides of ungalled leaflets. Al-Saghir and Porter (2005) claimed that all Pistacia species have no trichomes but others reported that their density

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

27

Fig. 7. Cross sections of the transition zone and more distal parts of the galled leaflet of S. betae gall types: (A and B) F1 gall type. (C and D) CLC gall type. (E–L) F2 gall type. The transition zone is characterised in the three gall types by small parenchyma cells (B, D, F). In the F1 and in the CLC galls there are differences in the length and structure between the two sides of the galled leaflet (A and C). (E) Section in a galled leaflet including the ostiolar zone, the transition zone and normal leaflet tissue. (F–G) detailed parts of section (E). Note that the transition zone is separated from the normal leaflet tissue by a medium-sized vein (G). (H) A short transition zone ending by the midvein. (I–K) A sharp transition from a gall to a leaflet tissue through a medium-sized vein. Note the change in cuticle thickness in the transition from gall to leaflet (J) and the sclereids that encircle 3/4 of the medium-sized vein (K). (L) Section of the most distal part of a F2 gall with the lamina attached to it. A rather long transition zone is separated from the normal leaflet tissue by a medium-sized vein. Abbreviations: cuticle (cu), gall (G), galled leaflet (GL), lumen (L), medium-sized vein (mdv), midvein (MIV), ostiole (os), ostiolar zone (OZ), sclereids (sc), transition zone (TZ). Scale bars: (A–H, J–L) = 200 ␮m. (I) = 1 mm.

is influenced by environmental factors (Ait Said et al., 2011; Belhadj et al., 2007). In Fordini galls, trichomes are not common and may be found mostly in the walls adjacent to the ostiole of F2 galls of Forda formicaria, on P. palaestina and P. terebinthus leaflets (Álvarez et al., 2009; Wool and Bar-El, 1995). They probably provide mechanical sealing for the gall entrance against intruders to F1 galls and then remain there when they turn into CLC galls. The cuticle of the external epidermis is thicker in the galls as compared with an intact leaflet epidermis. A thicker cuticle may reduce water loss through gall tissues and thus protect the aphids from desiccation or from fungal attacks. The possibility that the cuticle has an anti-desiccation role is supported by the fact that it was thinnest in the F1 galls that contain aphids only during the cooler and more humid spring and thickest in the F2 galls that

contain aphids throughout the hot and dry summer. The CLC galls exhibit an intermediate cuticle thickness but it should be noted that they were collected in June and therefore possibly did not reach their maximal thickness.

The modified parenchyma All galls show a remarkable hyperplasia and hypertrophy and a highly modified mesophyll that includes polyhedral-shaped parenchyma cells. Such modifications are common in aphidinduced galls (Álvarez et al., 2009; Wool and Bar-El, 1995) and other insect galls (Arduin et al., 2005; Kraus et al., 2002; Oliveira and Isaias, 2010), which are probably regulated by auxins and cytokinins (Yamaguchi et al., 2012). Similar modifications are well

28

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

known from leaf tissue cultures where very high levels of cytokinin and auxin are applied (Yancheva et al., 2003). The reddish-purple staining specifically in the gall’s outer parenchyma layers probably related to polyphenols which have a defensive role in aphid galls (Allison and Schultz, 1995; Rostás et al., 2013). Tannin levels in aphid galls increase with time (Álvarez et al., 2009; Álvarez, 2011b; Wool and Bar-El, 1995) which might explain the less prominent staining in short-lived F1 galls. The sclereid layer The most noticeable trait that distinguishes F2 and CLC galls from F1 galls and intact leaflets is the development of sclereids, specifically brachysclereids (“stone cells”). Brachysclereids commonly develop from parenchyma cells whose walls become secondarily thickened and lignified (Fahn, 1990). The formation of a lignified sclerenchyma sheath is common during the maturation of many seeds and insect galls (Oliveira and Isaias, 2010; Rohfritsch, 1992; Sliva and Shorthouse, 2006). So far it was found only in a single Fordini species, Forda marginata (Álvarez et al., 2009). The presence of a sclerenchyma layer represents an advanced state of differentiation and complexity and is an indication of the strong influence of the insect on the plant (Larew, 1982; Rohfritsch, 1992; Wells, 1921). The developmental signal for sclereid differentiation is not known, but sclereids often appear as a response to physiological disturbances such as wounding or aging (Fahn, 1990; Lev-Yadun, 1994). The sclereids provide the galls with protective mechanical strength and toughness against natural enemies (Stone et al., 2002) or provide structural physical supports (Dorchin et al., 2002). Sclereids are abundant in the proximal tip of the F2 galls (closest to the connection with the petiole) and in the midvein attached to it. They may physically strengthen the connection between the galled leaflet and the petiole and reduce the risk abscission. As a major component of the ostiolar zone of F2 and CLC galls, the sclereids possibly help blocking the entrance, making it harder for enemies and competitors to get in (Inbar, 1998; Wool, 2005). The vascular bundles and the midvein: Origin and implications on aphid nutrition The aphids feed directly from the phloem and therefore no specialised nutritive tissue is developed (Álvarez et al., 2009; Álvarez, 2011b). As in other leaflet-margin Fordini galls (Álvarez et al., 2009), all gall types of S. betae have a single layer of collateral veins (as in intact lamina) in their inner wall, where the phloem is oriented towards its external side, forcing the aphid to pass the xylem in order to feed. In the more complex Fordini midvein galls (induced by Geoica and Baizongia), the gall’s wall includes two layers of vascular bundles in which the phloem is oriented towards the inner cavity closer to the aphids (Álvarez, 2011b; Wool et al., 1999). Álvarez (2011a) demonstrated that this structure is derived from the leaflet’s midvein. The different structure of all S. betae gall types suggest that they originate from the leaflet lamina. Compared to the “midvein-based galls”, “lamina-based galls” (e.g., S. betae) induce small galls with weaker sinks, supporting fewer aphids (Burstein et al., 1994; Inbar et al., 2004). Within the F1 and CLC galls, it appears however, that the midvein is a better feeding site for the aphids than the other smaller veins. In these galls the midvein is often characterised by a whole (or almost whole) xylem ring, surrounded by a phloem ring; leaving part of the phloem in an orientation toward the lumen. Furthermore, the phloem fibres (in the CLC galls also the sclereids) are concentrated in the external side of midvein, away from the gall’s lumen. The schizogenous resin ducts that appear within the phloem (Sawidis et al., 2000) are also missing from the phloem facing the lumen. All these modifications should ease the access of the aphids to the phloem in the midvein region by reducing the mechanical

and chemical barriers near the sieve elements (see also Arduin et al., 2005). In addition, in the F2 gall type, the midvein is included only in the proximal side of the gall (closer to the petiole), usually as part of the ostiolar zone. Although the aphids in these galls probably do not feed directly from the midvein (but from the smaller veins in the gall) its proximity to the petiole may improve the ability of the galls to draw more resources (Whitham, 1978).

The role of aphid activity and galling site in controlling gall structure The three gall systems allow us to evaluate the relative role of the aphids and the plant (i.e., galling site) in controlling tissue modifications during gall formation. Characters that are shared only by F1 and CLC galls, such as gall shape, and formation of trichomes on the epidermis-lumen are not informative enough in this sense, since they can be the result of two hard-to-distinguish factors: (1) control by the fundatrix (i.e., its specific signalling or feeding behaviour) and/or (2) the galling site (leaflet midvein vs. margin) constraints (Wool, 1997). On the other hand, characters that are found only in F2 and CLC galls are controlled by the aphids and probably not by the site. The most pronounced example is the differentiation of sclereids. In addition is the accumulation of what seems like polyphenols in the outer parenchyma of the gall wall and cuticle thickening. It appears that these characters are not determined by the fundatrix, but rather depend on the continuous activity of the following generations in the gall (see also Álvarez et al., 2013). Sclereids started to occur only during May (L. Kurzfeld-Zexer, unpublished data). These observations concur with other studies that associated sclerenchyma differentiation with gall maturation and continuous insect activity (Rohfritsch, 1992). It is possible that signals provided by the fundatrix are not sufficient to induce sclereid differentiation. Nevertheless, the effect of a specific aphid generation on gall traits could be further explored by examining the galls through different developmental phases.

The evolution of three gall types in S. betae: Histologic viewpoint Wool and Burstein (1991) suggested that the primitive life cycle of S. betae include a single F1 gall. This scenario is supported by the fact that the entire life cycle can still be completed in CLC galls. This hypothesis is supported by the fact that the CLC gall displays a combination of histologic traits of both F1 and F2 gall (e.g., trichomes and sclereids, respectively) and no unique characters of its own. During the evolution of gall formation, traits that were under the fundatrix control and/or influenced by the gall location (e.g., gall shape and trichome formation) are maintained only in the F1 (and CLC) galls. On the other hand, traits whose induction requires the presence of the following generations (e.g., sclereids) are expressed in the F2 (and CLC) galls but were either lost or were never expressed in the F1 galls. Interestingly, CLC galls resemble final (F2) galls of S. betae but differ from most other Fordini final galls (see F. marginata galls in Álvarez et al., 2009). Inbar et al. (2004) suggested that the ancestral Fordini had a single (whole cycle) peashaped gall located on the leaflet midvein, like F1 galls. It is not known whether the ancestral gall type was similar to the CLC galls. The present study supports the concept of galls as an “extended phenotype” of the insect but within the constraints of the galling site. Furthermore, several gall traits depend on the continuous feeding activity and reproductions of the aphids.

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

Acknowledgments We thank Kamelia Loginovsky for her valuable advice on processing tissue samples. The assistance of Dr. Daniel Joel and Shira Gal with the section photography is greatly appreciated. We thank Goni Shavit and Anat Regev for the graphic design and anonymous reviewers for their helpful comments. This research was supported by The Israel Science Foundation (ISF grant No. 940/08 to M.I.).

References Ait Said, S., Fernandez, C., Greff, S., Derridj, A., Gauquelin, T., Mevy, J.P., 2011. Interpopulation variability of leaf morpho-anatomical and terpenoid patterns of Pistacia atlantica Desf. ssp. atlantica growing along an aridity gradient in Algeria. Flora 206, 397–405. Allison, S.D., Schultz, J.C., 1995. Biochemical responses of chestnut oak to a galling cynipid. J. Chem. Ecol. 31, 151–166. Aloni, R., Katz, D., Wool, D., 1989. Effect of the gall-forming aphid Slavum wertheimae on the differentiation of xylem in branches of Pistacia atlantica. Ann. Bot. 63, 373–375. Al-Saghir, M.G., Porter, D.M., 2005. Stomatal distribution in Pistacia sp. (Anacardiaceae). Int. J. Bot. 1, 183–187. Álvarez, N., 2011a. Initial stages in the formation of galls induced by Geoica utricularia in Pistacia terebinthus leaflets: origin of the two vascular bundles which. characterize the wall of the galls. AJPS 2, 175–179. Álvarez, R., 2011b. Microscopic study of the walls of galls induced by Geoica. utricularia and Baizongia pistaciae in Pistacia terebinthus: a contribution to the. phylogeny of Fordini. Arthropod-Plant Interact. 6, 137–145. Álvarez, R., Encina, A., Hidalgo, N.P., 2009. Histological aspects of three Pistacia terebinthus galls induced by three different aphids: Paracletus cimiciformis, Forda marginata and Forda formicaria. Plant Sci. 176, 303–314. Álvarez, R., González-Sierra, S., Candelas, A., Martinez, J.J.I., 2013. Histological. study of galls induced by aphids on leaves of Ulmus minor: Tetraneura ulmi induces globose galls and Eriosoma ulmi induces pseudogalls. Arthropod-Plant Interact. 7, 643–650. Ananthakrishnan, T.N., Raman, A., 1989. Thrips and gall dynamics. E.J. Brill Arch., The Netherlands, pp. 120. Ananthakrishnan, T.N., 1992. Unique aspects in the biology of thrips induced galls. In: Shorthouse, J.D., Rohfritsch, O. (Eds.), Biology of Insect-Induced Galls. Oxford University Press, New York, NY, pp. 185–195. Arduin, M., Fernandes, G.W., Kraus, J.E., 2005. Morphogenesis of galls induced by Baccharopelma dracunculifoliae (Hemiptera: Psyllidae) on Baccharis dracunculifolia (Asteraceae) leaves. Braz. J. Biol. 65, 559–571. Belhadj, S., Derridj, A., Aigouy, T., Gers, C., Gauquelin, T., Mevy, J.P., 2007. Comparative morphology of leaf epidermis in eight populations of atlas pistachio (Pistacia atlantica Desf., Anacardiaceae). Micros. Res. Tech. 70, 837–846. Bronner, R., 1992. The role of nutritive cells in the nutrition of cynipids and. cecidomyiids. In: Shorthouse, J.D., Rohfritsch, O. (Eds.), Biology of Insect-Induced Galls. Oxford University Press, New York, NY, pp. 118–140. Burstein, M., Wool, D., Eshel, A., 1994. Sink strength and clone size of sympatric gall. forming aphids. Eur. J. Entomol. 91, 57–61. Chakrabarti, S., 2007. Diversity and biosystematics of gall-inducing aphids (Hemiptera: Aphididae) and their galls in the Himalaya. Orient. Insects 41, 35–54. Cook, L.G., Gullan, P.J., 2008. Insect, not plant, determines gall morphology in the Apiomorpha pharetrata species-group (Hemiptera: Coccoidea). Aust. J. Entomol. 47, 51–57. Dawkins, R., 1982. The Extended Phenotype. Oxford University Press, Oxford. Dorchin, N., Freidberg, A., Aloni, R., 2002. Morphogenesis of stem gall tissues induced. by larvae of two cecidomyiid species (Diptera: Cecidomyiidae) on Suaeda monoica (Chenopodiaceae). Can. J. Bot. 80, 1141–1150. Dorchin, N., Hoffmann, J.H., Strike, W.A., Novák, O., Stranad, M., Van Staden, J., 2009. Sexually dimorphic gall structures correspond to differential phytohormone contents in male and female wasp larvae. Physiol. Entomol. 34, 359–369. Fahn, A., 1990. Plant Anatomy, fourth ed. Pergamon Press, Oxford. Gonc¸alves, S.J.M.R., Isaias, R.M.S., Vale, F.H.A., Fernandes, G.W., 2005. Sexual. dimorphism of Pseudotectococcus rolliniae Hodgson & Gonalves 2004 (Hemiptera Coccoidea Eriococcidae) influences gall morphology on Rollinia laurifolia Schltdl (Annonaceae). Trop. Zool. 18, 161–169. Gonc¸alves, S.J.M.R., Moreira, G.R.P., Isaias, R.M.S., 2009. A unique seasonal cycle in a leaf gall-inducing insect: the formation of stem galls for dormancy. J. Nat. Hist. 43, 843–854. Gullan, P.J., Miller, D.R., Cook, L.G., 2005. Gall-inducing scale insects (Hemiptera: Sternorrhyncha: Coccoidea). In: Raman, A., Schaefer, C.W., Withers, T.M. (Eds.), Biology, Ecology, and Evolution of Gall-inducing Arthropods. Science Publishers, Inc, Enfield, USA, Plymouth, UK, pp. 159–229. Inbar, M., 1998. Competition, territoriality and maternal defense in a gall-forming aphid. Ethol. Ecol. Evol. 10, 159–170. Inbar, M., Wink, M., Wool, D., 2004. The evolution of host plant manipulation by insects: molecular and ecological evidence from gall-forming aphids on Pistacia. Mol. Phylogenet. Evol. 32, 504–511.

29

Inbar, M., Kark, S., 2007. Gender-related developmental instability and herbivory of Pistacia atlantica across a steep environmental gradient. Folia Geobot. 42, 401–410. Kraus, J.E., Arduin, M., Venturelli, M., 2002. Anatomy and ontogenesis of. hymenopteran leaf galls of Struthanthus vulgaris Mart (Loranthaceae). Rev. Bras. Bot. 25, 449–458. Küster, E., 1911. Die Gallen der Pflanzen. S. Hirzel, Leipzig. Larew, H.G., 1982. A Comparative Anatomical Study of Galls Caused by the Major Cecidogenetic groups, with Special Emphasis on the Nutritive Tissue. Oregon State University, Corvallis, OR (Ph.D. Thesis). Lev-Yadun, S., 1994. Induction of sclereid differentiation in the pith of Arabidopsis thaliana (L.) Heynh. J. Exp. Bot. 45, 1845–1849. Mani, M.S., 1964. Ecology of Plant Galls. The Hague, The Netherlands, Walter Junk Publishers. Miller III, D.G., 1998. Life history, ecology and communal gall occupation in the manzanita leaf-gall aphid, Tamalia coweni (Cockerell) (Homoptera: Aphididae). J. Nat. Hist. 32, 351–366. Moran, N.A., 1988. The evolution of host-plant alternation in aphids: evidence for specialization as a dead end. Am. Nat. 132, 681–706. Nyman, T., Widmer, A., Roininen, H., 2000. Evolution of gall morphology and host–plant relationships in willow-feeding sawflies (Hymenoptera: Tenthredinidae). Evolution 54, 526–533. Oliveira, D.C., Isaias, R.M.S., 2010. Redifferentiation of leaflet tissues during midrib. gall development in Copaifera langsdorffii (Fabaceae). S. Afr. J. Bot. 76, 239–248. Price, P.W., Fernandes, G.W., Waring, G.L., 1987. Adaptive nature of insect galls. Environ. Entomol. 16, 15–24. Rhomberg, L., 1980. Causes of life history differences between the morphs of Pemphigus populitransversus. J. N.Y. Entomol. Soc. 88, 106–112. Raman, A., 1996. Nutritional diversity in gall-inducing insects and their evolutionary relationships with flowering plants. Int. J. Ecol. Environ. Sci. 22, 133–143. Raman, A., 2007. Insect-induced plant galls of India: unresolved questions. Curr. Sci. 92, 748–757. Rey, L.A., 1992. Development morphology of two types of Hymenopterous galls. In: Shorthouse, J.D., Rohfritsch, O. (Eds.), Biology of Insect-induced Galls. Oxford University Press, New York, NY, pp. 87–101. Rohfritsch, O., 1977. Ultrastructure of the nutritive tissue of the Chermes abietis L. fundatrix on Picea excelsa L. Marcellia 40, 135–150. Rohfritsch, O., 1992. Patterns in gall development. In: Shorthouse, J.D., Rohfritsch, O. (Eds.), Biology of Insect-induced Galls. Oxford University Press, New York, NY, pp. 60–86. Rostás, M., Maag, D., Ikegami, M., Inbar, M., 2013. Gall volatiles defend aphids against a browsing mammal. BMC Evol. Biol. 13, 193. Ruzin, S.E., 1999. Plant Microtechnique and Microscopy. Oxford University Press, New York. NY. Sawidis, T., Dafnis, S., Weryzko-Chemielewska, E., 2000. Distribution, development. and structure of resin ducts in Pistacia lentiscus var. chia Duhamel. Flora 195, 83–94. Shorthouse, J.D., Rohfritsch, O., 1992. Biology of Insect-induced Galls. Oxford University Press, New York, NY. Sliva, M.D., Shorthouse, J.D., 2006. Comparison of the development of stem galls induced by Aulacidea hieracii (Hymenoptera: Cynipidae) on hawkweed and by Diplolepis spinosa (Hymenoptera: Cynipidae) on rose. Can. J. Bot. 84, 1052–1074. Stern, D.L., 1995. Phylogenetic evidence that aphids, rather than plants, determine gall. morphology. Proc. R. Soc. Lond. B 260, 85–89. Stone, G.N., Cook, J.M., 1998. The structure of cynipid oak galls: patterns in the evolution of an extended phenotype. Proc. R. Soc. Lond. B 265, 979–988. Stone, G.N., Schönrogge, K., Atkinson, R.J., Bellido, D., Pujade-Villar, J., 2002. The population biology of oak gall wasps (Hymenoptera: Cynipidae). Annu. Rev. Entomol. 47, 633–668. Weis, A.E., Walton, R., Crego, C.L., 1988. Reactive plant tissue sites and the population biology of gall makers. Annu. Rev. Entomol. 33, 467–486. Wells, B.W., 1921. Evolution of zoocecidia. Bot. Gaz. 71, 358–377. Wertheim, G., 1954. Studies on the biology and ecology of the gall producing aphids of the tribe Fordini (Homoptera Aphidoidea) in Israel. Trans. Entomol. Soc. Lond. 105, 79–96. Whitham, T.G., 1978. Habitat selection by Pemphigus aphids in response to resource limitation and competition. Ecology 59, 1164–1176. Wool, D., 1995. Aphid induced galls on Pistacia in the natural forest of Israel: which, where, and how many? Isr. J. Zool. 41, 591–600. Wool, D., 1997. The shapes of insect galls: insect control, plant constraints and phylogeny. In: Raman, A. (Ed.), Ecology and evolution of plant-feeding insects in natural and manmade environments. International Scientific Publications, New Delhi, pp. 203–212. Wool, D., 2004. Galling aphids: specialization, biological complexity, and variation. Annu. Rev. Entomol. 49, 175–192. Wool, D., 2005. Gall-inducing aphids: biology ecology and evolution. In: Raman, A., Schaefer, C.W., Withers, T.M. (Eds.), Biology, ecology and evolution of gallinducing arthropods. Science Publishers Inc., Enfield, USA, Plymouth, UK, pp. 73–132. Wool, D., Bar-El, N., 1995. Population ecology of the galling aphid Forda formicaria Von Heyden in Israel: abundance, demography and gall structure. Isr. J. Zool. 41, 175–192.

30

L. Kurzfeld-Zexer et al. / Flora 210 (2015) 19–30

Wool, D., Burstein, M., 1991. A galling aphid with extra life-cycle complexity: population ecology and evolutionary considerations. Res. Popul. Ecol. 33, 307–322. Wool, D., Aloni, R., Ben-Zvi, O., Wollberg, M., 1999. A galling aphid furnishes its home with a built-in pipeline to the host food supply. Entomol. Exp. Appl. 91, 183–186. Yamaguchi, H., Tanaka, H., Hasegawa, M., Tokuda, M., Asami, T., Suzuki, Y., 2012. Phytohormones and willow gall induction by a gall-inducing sawfly. New Phytol. 196, 586–595.

Yancheva, S.D., Golubowicz, S., Fisher, E., Lev-Yadun, S., Flaishman, M.A., 2003. Auxin type and timing of application determine the activation of the developmental. program during in vitro organogenesis in apple. Plant Sci. 165, 299–309. Zohary, M., 1952. A monographical study of the genus Pistacia. Palestine J. Bot. Jerusalem Ser. 5, 187–228.