European and Mediterranean Myzocallidini Aphid Species: DNA Barcoding and Remarks on Ecology with Taxonomic Modifications in An Integrated Framework

Simple Summary Myzocallis is a Holarctic genus of monoecious species of aphids mostly hosted by plants belonging to Fagales. To date, extensively morphological studies have been carried out on this group of aphids, but only sporadic molecular studies have been performed to understand the relationships among the different species. With the aim improve knowledges on these aspects, almost all species of the European and Mediterranean Myzocallidini species were investigated. As a consequence of the results obtained, Myzocallis (Agrioaphis) leclanti originally described as a subspecies of M. (A.) castanicola and M. (M.) schreiberi, considered a subspecies of M. (M.) boerneri, should be regarded at a rank of a full species. Moreover, the subgenus Agrioaphis, Lineomyzocallis, Neomyzocallis and Pasekia were elevated to the rank of genus, while Myzocallis remain as such. Abstract The genus Myzocallis Passerini (Hemiptera, Aphididae, Calaphidinae, Myzocallidini) is a rather primitive group of aphids currently comprising 45 species and 3 subspecies, subdivided into ten subgenera, three of them having a West Palaearctic distribution. The majority of the species inhabit Fagales plants and some of them are considered pests. Despite their ecological interest and the presence of some taxonomic controversies, there are only a few molecular studies on the group. Here, the main aims were to develop a DNA barcodes library for the molecular identification of West Palaearctic Myzocallis species, to evaluate the congruence among their morphological, ecological and DNA-based delimitation, and verify the congruence of the subgeneric subdivision presently adopted by comparing the results with those obtained for other Panaphidini species. These study findings indicate that Myzocallis (Agrioaphis) leclanti, originally described as a subspecies of M. (A.) castanicola and M. (M.) schreiberi, considered as a subspecies of M. (M.) boerneri, should be regarded at a rank of full species, and the subgenera Agrioaphis, Lineomyzocallis, Neomyzocallis, Pasekia were elevated to the rank of genus, while Myzocallis remain as such.


Introduction
The Holarctic genus Myzocallis Passerini (Hemiptera, Aphididae, Calaphidinae, Myzocallidini) currently lists 45 species, including a few subspecies [1,2]. It is subdivided into ten subgenera, three of them having a West Palaearctic native area of distribution as quoted in the revision of all species-group on a worldwide scale [3,4]. Except for the Nearctic M. asclepiadis (Monell) living on Asclepias (Apocynaceae), all other Myzocallis genus, benefitting also of comparison with molecular data of other Panaphidini species; and (iii) verify the congruence of the subgeneric subdivision presently adopted.

Aphids Collection and Specimen Identification
The study was carried out on about 400 samples representatives of almost all Myzocallis species present in the Mediterranean area plus other closely related Panaphidini species (Apulicallis trojanae Barbagallo & Patti, Tuberculatus eggleri Börner, T. neglectus Krzywiec, T. remaudierei Nieto Nafría, Hoplocallis ruperti Pintera and H. picta Ferrari, Siculaphis vittoriensis Quednau & Barbagallo). The specimens were dropped on the surface of a wooden plate by beating the leaves and then placed in tubes containing 85% alcohol. The leaves and acorns of each plant from which the aphids were collected were taken and stored for subsequent recognition. The host plants on which the analysed individuals were collected were identified following most accredited European Flora handbooks [30,31]. Metadata on the analysed specimens, such as host plants, geographical coordinates, sampling date and GenBank accession, are reported in Table S1.
For each sample collected, a preliminary classification of individuals was carried out through observation under the microscope. Subsequently, to confirm the first visual morphological classification a number of specimens were mounted on slides, according to the current preparation method for aphids and particularly for the softest Calaphidinae species [32,33]. The identification was carried out using characters reported in the key [4] and the comparison to specimens in the collection of senior co-authors (S.B. and J.M.N.N.). One single alate viviparous has been seen for M. persica Quednau & Remaudière and also M. taurica Quednau & Remaudière; moreover, few specimens of M. glandulosa Hille Ris Lambers (either as alate and apterous viviparous females) have been seen, thanks to the courtesy of Mr. Paul Brown-The Natural History Museum of London. Slides are available in the co-authors' collections (Dept. of Agri-Food and Environmental Systems Management, University of Catania, Italy and the Dept. of Biodiversity and Environmental Management, Univ. of León, Spain). The largest part of the collected specimens was then stored in 95% (−20 • C) ethanol for the subsequent molecular analyses.

DNA Extraction, Amplification and Sequencing
After the morphological examination under the stereo-microscope, 3-4 samples, representative of each sample, were randomly selected for the molecular analysis. Total DNA was extracted from single individuals using the DNeasy Blood & Tissue kit (Qiagen, Hilden, Germany) following the instruction suggested by the manufacturing company. The nondestructive method [10] was used, so to not preclude subsequent morphological analyses of the specimen if molecular analysis makes it necessary. A fragment of mitochondrial COI was amplified using the universal primers LCO1490 and HCO2198 [34]. All PCRs were performed in 10 µL, with 4.25 µL buffer premix 2 × F (FailSafe tm PCR Premix Selection Kit, Epicentre Technologies, Thane, India), 0.5 µL of each primer (10 pmol, 0.25 µL Taq polymerase (Life Technologies, Carlsbad, CA, USA) and 2 µL DNA template. Thermal PCR cycle and electrophoresis conditions see [25]. PCR products were sent for sequencing to BMR Genomics (Padua, Italy) using ABI PRISM 3730XL DNA sequencer. All chromatograms were evaluated using 4Peaks [35], low-quality sequences were excluded from the following analyses and doubtful initial or final regions were pruned. The presence of open reading frame was assessed in order exclude nuclear mitochondrial pseudogenes. The developed sequences were deposited in Genbank (accessions list in Table S1).

Nucleotide Distance Analyses and Taxa Molecular Delimitation
All COI sequences of the Panaphidini subfamily publicly available on BOLD [36] were retrieved, and then aligned together with those developed in this study. Alignment was performed using MUSCLE algorithm [37] implemented in MEGA X [38]. Alignment was trimmed to retain the region shared among the majority of the sequences (616 bp segment within Folmer region [34] and then all the sequences with length <390 bp were excluded. R library Haplotypes (https://biolsystematics.wordpress.com/r/ accessed on 18 March 2022) was used for reducing haplotypes within each species. Finally, some of the sequences retrieved from BOLD were removed from the dataset or their identifier modified for the following reasons: incompatibility between the individual morphological identification and the collection host plant; sequences related to misidentifications (previously signalled by other scientific works) (Table S2). This dataset was then split into two sub-datasets, i.e., Myzocallis genus sub-dataset and a dataset including all Panaphidini sequences except for Myzocallis ones, in order to perform the analyses described hereafter.
From the Panaphidini dataset, the sequences belonging to the Myzocallis genus were extracted using R software and a Kimura-two parameter (K2P) [39] pairwise nucleotide distance matrix was estimated starting from them using ape R library [40]. The obtained nucleotide distance matrix was analysed for extrapolating summary statistics on intraspecific and interspecific distances, intrasubgeneric and intersubgeneric distances, and intrageneric distances using the R library spider [41].
Species delimitation analyses were performed on the same Myzocallis nucleotide sequences dataset using two species delimitation methods (i) 2% nucleotide distance threshold as species clustering threshold, a value corresponding to the maximum intraspecific distance estimated by [23] on aphids COI sequences. This analysis was carried out using the R package spider [41]. (ii) Assemble Species by Automatic Partitioning Estimation (ASAP) [42]. The delimitation was performed on the server (https: //bioinfo.mnhn.fr/abi/public/asap) using the K2P model [39] and the remaining parameters were set as default. ASAP delimitation was defined by evaluating both the partitions with the first and the second best asap-score.
The previously aligned COI gene sequences of the genus Myzocallis were used to infer a single-gene phylogenetic tree using both Bayesian and Maximum Likelihood approaches. The best nucleotide substitution model was estimated using PartitionFinder2 [43] and selected according to the Bayesian information criterion (BIC) [44]. According to BIC, the best model of nucleotide substitution resulted the HKY model [45] with gamma distribution (Γ) and proportion of invariable sites (I). Bayesian inference was performed using MrBayes 3.2.2 [46] with two independent runs of 3 × 10 7 generations (sample frequencies: every 100 generations; stationarity reached when the average standard deviation of split frequencies <0.01) and the nucleotide substitution model settled according to the results of the model selection analysis. The convergence of the runs was visually inspected using TRACER [47] and an appropriate number of sampled trees were discarded as burn-in. The Maximum Likelihood inference was performed using PhyML 3.0 [48] implementing: the selected model of nucleotide substitutions; tree searching operations accounting for the best between the nearest neighbour interchange and the subtree pruning and regrafting; approximate Likelihood-Ratio Test (aLRT) [49] as branch support.
The sub-dataset of Panaphidini sequences not including Myzocallis genus was created excluding the sequences of the Myzocallis genus from the full Panaphidini dataset using R software. This dataset was used for estimating a K2P pairwise nucleotide distance matrix from which intergeneric distances were derived. From the same dataset, the sequences of genera represented from at least two species were extracted and used for the estimation of K2P intrageneric distances.

Results
In this study, a DNA barcode library including sequences from 63 Panaphidini individuals was developed. Processed individuals belonged to 19 species, 13 of them of the genus Myzocallis, while the remaining of the genera Tuberculatus (three species), Hoplocallis (one species), Apulicallis (one species) and Siculaphis (one species) ( Table S1). The mean length of the obtained barcode sequences was of 616 bp [range: 600-616 bp] with the following average base composition: A = 34.6%, C = 15.1%, T = 40.5%, G = 9.8%.

Nucleotide Distance Analyses
The Myzocallis COI sequences developed in this study plus all the sequences available for this genus in BOLD, for a total of 90 sequences belonging to 17 species (representative of five different Myzocallis subgenera, i.e., M. Agrioaphis, M. Lineomyzocallis, M. Myzocallis, M. Neomyzocallis, M. Pasekia) were assembled in a dataset that includes~43% of the species currently described for this genus.
Nucleotide divergence within and among Myzocallis species resulted in a mean value of 2.1% (range: 0.2-11.4%) and 10.4% (range: 0-18.3%) respectively, while intrageneric divergence was estimated to be in mean of 9.6% (Figure 1a). Myzocallis subgenera intraand inter-subgeneric nucleotide distances resulted to be in mean of 4.5% and 11.6%, respectively (Figure 1b  The dataset assembled in this study including all sequences of Panaphidini except for Myzocallis (de novo developed plus BOLD database publicly available barcodes) resulted to be composed of 580 COI sequences. Specifically, it included 118 species belonging to 41 genera with a mean intergeneric divergence between genera of 13.1% ( Figure 1g) and mean intrageneric divergence, estimated only on genera represented by at least two species (520 sequences of 20 genera), of 6.9% (Figure 1f).

Species Delimitation of the Genus Myzocallis
The species delimitation analyses performed on the genus Myzocallis dataset using two molecular delimitation methods (i.e., the 2% clustering threshold and ASAP) produced comparable quite results and partially reflect the classical subdivision based on morphology. Specifically, 2% threshold delimitated 21 evolutionary units within the dataset, 11 of them exactly matched the morphological species. The species M. walshii and M. castanicola were both split in three evolutionary units, while all M. carpini and M. coryli were merged in the same one (except for two samples). Finally, also M. asclepiadis and M. punctata were merged in a single unit (Figure 2). Species delimitation adopting ASAP method led to almost identical results with the exceptions of M. carpini and M. coryli whose sequences were split into different evolutionary units but never merged together, and of M. castanicola, whose sequences were split in four units, one of them including only M. leclanti sequences (Figure 2). In the Myzocallis genus dendrogram, inferred using the same COI sequences dataset, not all the species resulted monophyletic, with incongruences between specimens' morphological identification and monophyly in tree involving some of the species already highlighted from species delimitation analyses (six species, Figure 2). Whereas, all Myzocallis subgenera, except for Agrioaphis, formed monophyletic clusters (nodes support aLRT ≥ 0.85, except for Pasekia; Figure 2). Noteworthy, the two groups identified within the subgenus Myzocallis from nucleotide distance analyses (i.e., subgen Myzocallis group A and Myzocallis group B) resulted in two well supported monophyletic groups (aLRT = 0.95 node support) (Figure 2).

Discussion
The analyses performed in this study shed light on some aspects of the taxonomy of the genus Myzocallis. The main taxonomic inferences derived from this study can be summarized as follow: (i) two different groups are present within the subgenus Myzocallis whose taxonomic rank should be better investigated; (ii) the nominal taxon M. castanicola leclanti Quednau & Remaudière possibly should be elevated to the level of full species; (iii) the status of valid species of M. schreiberi Hille Ris Lambers & Stroyan, which in the past was considered as a synonym of M. boerneri Stroyan, has been confirmed also based on molecular analyses; (iv) the taxonomic hierarchy of Myzocallis subgenera have to be revised, together with other related genera of Myzocallidini.

Myzocallis Subgenus Myzocallis
The analyses revealed that the subgenus Myzocallis is divided into two well separated clusters (in terms of nucleotide distance) based on COI gene sequences. Myzocallis group A showed a very low intraspecific nucleotide divergence based on COI, independently from their geographical origin (except for one sequence of M. carpini that will be further discussed). M. coryli was morphologically well-studied by several authors, but the presence of morphological variation among the populations of this species was never reported. However, for this species an unusually high intraspecific genetic variability was already observed (on the COI gene), especially between individuals from different geographical areas [10,50]. In previous researches, the authors hypothesized that M. coryli is a cryptic species complex. The nucleotide sequences of M. coryli analysed in this study (18 sequences in total, 14 of which were mined from BOLD) resulted subdivided into three clusters plus one independent sequence in the COI dendrogram and were differentially delimited by the two species delimitation methods adopted, i.e., two (clustering threshold) and three (ASAP) different evolutionary units ( Figure 2). Intraspecific nucleotide distance of such a species ranged from 0.02% to 5%. Based on this evidence, M. coryli may represent a complex of cryptic species. Anyway, since M. coryli is a holocyclic and monoecious species on Corylus genus, the nucleotide distances and clusters observed could be explained by the wide geographic distribution of its host plant. Although the latter has a European-Caucasian origin, currently it is widespread throughout the world for agricultural production with M. coryli that followed its distribution. This may have determined a progressive adaptation of the aphid populations to the different environmental conditions and led to some genetic divergence between them, without evolving morphological differences and maybe nether incurring in speciation. This situation might have some analogy B. helichrysi, where a COI nucleotide distance of 2.7% was detected between two different populations and for which the existence of two cryptic species has been hypothesized [51,52]. However, further investigation within this group should be undertaken even with additional mitochondrial and nuclear markers.
The analyses performed in this study clearly distinguished the holm oak aphid M. (M.) schreiberi from the Turkey oak aphid M. (M.) boerneri Stroyan. Both species delimitation methods recognised them as separated species, but also, they have fallen into separate clusters in COI tree (Myzocallis group A, Myzocallis group B) (Figure 2). The two species are morphologically very similar, so in the past, they have often been confused, doubting whether they were distinct species or assuming that M. schreiberi was an anholocyclic form of M. boerneri [3,4,53]. This misunderstanding is probably at the origin of the incorrect classification of the COI sequences ACEA810-14 and GBMHH5717-14 available on BOLD (renamed as M. schreiberi for these study analyses; see Materials and Methods). While describing M. (M.) schreiberi from specimens collected on Q. ilex in Italy and England, Hille Ris Lambers & Stroyan [54] listed the morphological characters distinguishing this species from other European congeneric taxa known at the time. While the identification keys to distinguish M. schreiberi and M. boerneri are provided by Barbagallo & Massimino Cocuzza [11]. These two aphids have a rather different host plant preference. M. schreiberi usually lives on Q. ilex and on other evergreen-leaved oak species, such as Q. suber and Q. x crenata. All these oaks have coriaceous leaves with a grey-tomentose texture beneath, to which this aphid species is likely adapted to feed on through its rather acute shaped last rostral joint. In the present study, no genetic difference was found among the populations of M. schreiberi living on Q. ilex, Q. suber and Q. x crenata. M. boerneri is widely distributed in the West Palaearctic region (Europe, Middle East) and its main host plant is the deciduous Q. cerris.
The M. (M.) glandulosa Hille Ris Lambers [55] and M. (M.) occidentalis Remaudière and Nieto Nafria [56] are both characterized by a rather long last rostral joint. The former was described from Q. ithaburensis in the Middle East [48]. Several years later, the alate viviparous female of M. (M.) glandulosa was figured and Q. persica was added as a further host plant for this species [4]. The individuals whose COI sequences were analysed in the present study were collected in Israel from the type locality. M. (M.) occidentalis has been described from specimens collected in South-Western Europe (France, Spain); the only host plant known is Q. pyrenaica, on which the aphid performs a holocyclic life cycle [8,55]. Also included in this group is M. macrolepidis, a new species of recent description [11]. Moreover, the sequence labelled as M. occidentalis presents on NCBI (accession number GBMIN66582) obtained from a specimen collected in California on Quercus sp. [10] should be regarded as belonging to a different taxon and would deserve in-depth analysis. Finally, also the classification as M. occidentalis of the samples caught with suction or yellow water traps in Serbia and Greece could be incorrect [57,58].

Myzocallis Subgenus Agrioaphis
The subgenus Agrioaphis Walker is represented so far by two species Myzocallis myricae and M. castanicola (the latter including two subspecies, M. castanicola castanicola and M. castanicola leclanti) [3]. M. castanicola mainly lives on Quercus, but it was recorded as well from Castanea by several authors. The very common populations living on C. sativa in southern Europe and the Middle East, usually belong to the subspecies M. castanicola leclanti. On chestnut it performs a monophagous and holocyclic life cycle and not rarely is considered a noxious aphid species. In Italy, M. castanicola castanicola is quite common in northern areas of the peninsula, mainly on Q. petraea and its hybrids x Q. pubescens (group) and unlikely it can be found there on different Quercus-species (including Q. robur) or on C. sativa, unless perhaps as occasional vagrant alates. In Spain, the aphid commonly lives on Q. pyrenaica and it is occasionally detected on additional oak species [8].
Based on the results obtained in the present study, the bog myrtle aphid M. myricae (Kaltenbach) is clearly distinguishable from M. castanicola. Interestingly, M. castanicola was split into three or four evolutionary units based on the delimitation methods adopted (Figure 2). These units appear to be related with the host plant on which individuals were collected. In the case of ASAP, the evolutionary units were composed as follow (i) individuals collected from Q. pyrenaica and Q. cerris (plus two sequences mined from BOLD for which the host plants are not specified), (ii) individuals collected from Q. petraea in North Est Italy, (iii) individuals collected on Castanea sativa and currently classified as M. castanicola subsp. leclanti, (vi) a sole individual whose sequence was mined from BOLD (GBMHH16898-19). A 2% delimitation threshold partially confirmed these results, but units (ii) and (iii) were merged together since the nucleotide distance between individuals is~2%. Further analyses are needed to shed light on the taxonomic status of the individuals following in these clusters, but we hypothesis M. castanicola GBMHH16898-19 is a misidentification (actually this sequence may belong to a species not present in the dataset of this study). On the other hand, the analyses here performed clearly discriminate M. castanicola castanicola and M. castanicola leclanti, as belonging to two separate evolutionary units. M. castanicola castanicola and M. castanicola leclanti also clearly differ from the morphological point of view. Specifically, M. castanicola leclanti has (i) paler yellow colour of nymphs, alate viviparous females and sexuales; (ii) paler and less extensive dorsal sclerified areas, particularly those on abdomen; (iii) more haired last rostral joint. Quednau & Remaudière [3] and Quednau, [4] provided the morphological description and illustration of M. castanicola leclanti, adding the comparative range of their variation within the different seasonal morphs of M. castanicola castanicola.

Myzocallis Subgenus Pasekia
The subgenus Pasekia Aizenberg is represented so far by five species, i.e., Myzocallis persica, Myzocallis taurica, Myzocallis komareki, Myzocallis mediterranea and Myzocallis cocciferina. All of them are well studied from the morphological point of view and a key for their identification (as viviparous alates and nymphs, including males) has been reported [3,4].
M. (Pasekia) persica Quednau & Remaudière and M. (P.) taurica Quednau & Remaudière are apparently confined to eastern Mediterranean. Both species are holocyclic, the former on Q. persica, and the latter on Q. coccifera (maybe also other oak species). Unluckily, these species were not included in the present study since no specimen was available for the molecular analyses.
M. (P.) komareki Pašek has a wider distribution, it was recorded from Central Europe southward to the Mediterranean and the Middle East on several oak species, as well as on Castanea sativa [1,3,8,59]. The nucleotide distance analysis of this study showed low intraspecific variability (<2.1%) for M. komareki, despite the analysed specimens being collected from different oak species and various geographical localities. It can be argued that, due to its polyphagy, this species can easily adapt to the different oak species, but without occurring in isolation of its populations.
M. (P.) mediterranea Quednau & Remaudière is recorded so far from France, Spain and Italy, and probably it is widespread also in other Mediterranean countries. The species inhabits several oaks such as Q. pubescens and strictly allied taxa (Q. congesta, Q. virgiliana and Q. dalechampii) [60] and only occasionally it has been collected on oaks of different groups (i.e., Q. ilex). In the past, this aphid was frequently confused with M. (P.) komareki, because of their morphological similarity [61,62]. Usually, M. (P.) mediterranea develops through a regular holocycle. Nevertheless, in more warm habitats (such as in southern Italy) the amphigonic morphs appear very late in the season (from the end of December to the first half of February), inhabiting the host plant leaves still green in Q. pubescens-group species (semi-evergreen oaks). In such a case, part of the aphid population can overwinter anholocyclically on these oaks and the survived specimens move to new oak blossoms by the end of March or early in April [63] (as M. komareki). Also, in the case of M. mediterranea, a low intraspecific variability (<1% nucleotide distance) was found between specimens considered in this study, however all of them were collected from a limited geographic area (Sicily) on Q. pubescens.
Finally, M. (P.) cocciferina Quednau & Barbagallo is present in southern Europe and Mediterranean countries, from Spain eastward to Lebanon on Q. coccifera s. lat. and oc-casionally on Q. ilex or their hybrids (ilex x coccifera). In Spain, the aphid is recorded on Q. ilex and reported as well for Portugal and North Africa [8]. This species was originally attributed [64] to subgenus Agrioaphis sensu Richards (1968) and later transferred to Pasekia [3,64]. In the present study, molecular and morphological species delimitation were consistent for M. cocciferina. A nucleotide distance of~1.5% was estimated between specimens collected from Q. coccifera in Sicily and Apulia and those collected from Q. ilex or its hybrid x Q. coccifera (postulated as Q. soluntina Tineo ex Lojacono by Giardina et al. [65]).

Nearctic Myzocallis Subgenera Lineomyzocallis and Neomyzocallis
In this study, a few specimens belonging to the Nearctic subgenera Lineomyzocallis (M. walshii, M. bellus and M. ephemerata) and Neomyzocallis (M. asclepiadis and M. punctata) were included. Here, two relevant situations were observed: (i) a high intraspecific variability of M. walshii (range intraspecific nucleotide distance: 0.2-8%); (ii) a very low nucleotide distance between M. asclepiadis and M. punctata (<0.2%). All the species delimitation methods adopted split M. walshii into three evolutionary units, two of them including one individual each ( Figure 2). The black-bordered oak aphid, M. walshii lives on Quercus of the group of red oaks (Quercus section Lobatae) and is native to North America, where it is widely distributed. In 1989 the species was found for the first time in France and in about twenty years it spread throughout the European continent [1]. None of the numerous studies performed in the last decades on M. walshii reported the presence of intraspecific morphological or biological variability. Possibly, the high genetic divergence observed in this study could be related to specimen misidentifications, rather than to the existence of cryptic diversity. Both the molecular species delimitation methods used in this study delimited M. asclepiadis and M. punctata in a single species, in accordance with what was already observed [10], i.e., some sequences present on BOLD may have been mistakenly attributed to M. asclepiadis.

Conclusions
Systematics is a hierarchic science fundamentally built on the relationships of affinity between species [66]. COI is a DNA marker known to be effective in discriminating insect taxa at the lowest taxonomic levels. Combining molecular (COI gene), morphological and ecological information in an integrative framework makes possible to resolve numerous taxonomic issues concerning insect species [67][68][69][70][71]. In particular, the most accurate results are obtained when expert morphologists on the target group are involved in the evaluation of the signal resulting from molecular analyses, in this case is also easier to distinguish between extrinsic and intrinsic errors related to the species molecular identification and delimitation [71][72][73]). For some insect groups, as in the case of aphids, the evaluation of the ecological information is important in the process of species identification and delimitation as well. Hence the relevance of coupling DNA sequences with specimen's collection locality and date, and further ecological information if present (e.g., collection habitat or host plant), even in the phase of molecular data publication.
However, also in this context, the use of integrative taxonomy (and COI as DNA marker) to define other taxonomic categories than species could sound as improper. Whereas, in previous studies it was done and valuable results were obtained. For example, the COI nucleotide distances have been examined for evaluating the range of variation between genera and subfamilies of Aphididae [10,23,27]. Lee et al. [10] developed the barcode sequences for 154 Aphididae species (72 genera, 11 subfamilies), and reported a mean pairwise divergence between specimens of different genera of 8.9% (range from 1.6 to 19%). Other studies, on Adelgidae and Eulachnini aphids [23], found a mean intergeneric distance of 9.6% and 11.7%, respectively. In a further investigation, a mean intergeneric distance of 7.7% (from 5 to 9.7%) in the tribe of Macrosiphini and 10.4 (range 8.9% to 12.4%) in Aphidini were found [27].
The Myzocallis subgenera analysed in this work are morphologically distinguishable through characters of well-verified validity [4], recognized as adequate by aphid-taxonomists. In the present study, a substantial nucleotide distance was found between the Myzocallis subgenera (Figure 1), reinforcing the validity of these taxa, regardless of their hierarchical position. The nucleotide distance values are sufficient to justify the elevation of the subgenera to the rank of genus. In particular, considering that the inter-subgeneric distances estimated in this study for Myzocallis subgenera are comparable to those observed between genera of the Panaphidini tribe (Figure 1), that certainly include species phylogenetically more distant than those of the Myzocallis genus.
Several authors stated that different genera have to be monophyletic [66,74], and the conclusion drawn by the present work meets this condition. Furthermore, this evidence could be supported also from the morphological point of view. Some of the most important morphological characters [3,4] (i.e., spinal hairs on abdominal tergites, chaetotaxy of immature morphs and oviparae) to distinguish Myzocallis subgenera from each other, are the same that were used to distinguish the genera Hoplocallis (until a few years ago considered as a subgenus of Myzocallis), Apulicallis and Siculaphis.
Consequently, we elevated these subgenera to the rank of genus: Agrioaphis Walker, stat. n. [