Types Holotype queen labeled'Brazil: Paraopeba MG/ 06 Oct 2003/ D. J. Souza'(MZUSP). Measurements (in mm): HL = 1.5; HW = 1.4; ML = 0.7; WL = 2.6; SL = 1.5; ED = 0.4.
Derivation of specific name This species is named after Amelia Maria de Souza, mother of the first author of this work.
Description The measurements obtained from 20 queens are presented in Table 1. This new species has a palpal formula of 4, 2 and 11 antennal segments as is typical for attine ants. Acromyrmex ameliae queens have large and convex eyes and the inferior pronotal spines are straight and forward-positioned as in the host species. The color of the analyzed parasite queens vary from brownish to brownish-black and this is not associated with the subspecies of host ants since one can find queens of the parasite of either color in a single nest. Worker color is the only character that is presently used to separate the two host subspecies: A. subterraneus subterraneus and A. subterrraneus brunneus . The latter has a brownish-black color whereas the former is light brown to yellow (Gon9alves, 1961). The queens of A. ameliae are much smaller than those of its hosts, with a WL ~ 0.6 as great (Table 1). Acromyrmex ameliae further has a more abundant pilosity with thicker and longer hairs on the gaster, on the dorsal portion of the alitrunk and on the anterior portion of the head, in comparison to that on the host subspecies. The parasite queen also has prominent ridges on the head (Figs. 1 and 2) and on the first segment of the gaster and expansions on the anteroventral margin of the postpetiole that are not seen on the host subspecies (Figs. 5 and 6). The tubercles on the gaster of A. ameliae are more or less ordinated in four longitudinal lines similar to the host species. However, these tubercles are very much reduced and less prominent when compared to those of the hosts.
HLHWWLQueensA. subterraneus subterraneus2.1 ± 0.12.2 ± 0.24.0 ± 0.1(n = 9, 6 nests)A. subterraneus brunneus2.0 ± 0.12.1 ± 0.14.0 ± 0.1(n = 8, 8 nests)A. ameliae1.3 ± 0.11.3 ± 0.12.4 ± 0.1(n = 20, 4 nests)MalesA. subterraneus subterraneus1.3 ± 0.11.6 ± 0.13.1 ± 0.2(n =8, 7 nests)A. ameliae1.0 ± 0.11.2 ± 0.12.3 ± 0.7(n =15, 4 nests)WorkersA. subterraneus subterraneus0.9 ± 0.11.1 ± 0.11.5 ± 0.1(only minor workers)(n = 16, 3 nests)A. ameliae0.9 ± 0.31.1 ± 0.31.5 ± 0.1(n = 30, 3 nests)
Paratype male labeled 'Brazil: Paraopeba MG/ 06 Oct 2003/ D. J. Souza' (MZUSP). Measurements (in mm): HL = 1.0; HW = 0.9; ML = 0.6; WL = 2.2; SL = 1.3; ED = 0.4.
Description The measurements of 15 males are presented in Table 1. The males of A. ameliae have 13 antennal segments. This characteristic was not constant since five individuals seemed to have 12 segments as a consequence of the fusion of segments 4 and 5 of the antennal funiculum like in the host males (compare Figs. 3 and 4). This fact was also observed by Schultz et al. (1998) in A. insinuator . Males of A. ameliae are visually smaller than the males of the studied host species (about 1.2 times). The antenna has a color gradient which ranges from dark brown to dark yellow when going from tip to base. The color of the males as well as of the queens is close to that of the host subspecies A. subterraneus subterraneus , that is, very dark brown independent of the parasitized subspecies. However, as pointed out by Goncalves (1961), the character color shows variation even inside the same nest. Newly emerged males and queens of A. ameliae of a lighter color were observed by the authors in the nests, but they became dark brown after a few days had elapsed. Their mandibles have a terminal tooth greater than the other teeth, which vary from 5 to 7 (Fig. 3). The ventral portion of the post-petiole of the parasite has irregular projections that are not seen in the host whose petiole margin is more regular and presents a concavity not observed on the parasite (Figs. 7 and 8). Ridges and tubercles can be observed on the gaster of A. ameliae males, but these are missing in the host species whose gaster is smooth and shiny.
Paratype worker labeled 'Brazil: Paraopeba MG/ 06 Oct 2003/ D. J. Souza' (MZUSP). Measurements (in mm): HL = 0.6; HW = 0.7; ML = 0.3; WL = 0.9; SL = 0.8; ED = 0.1.
Description We verified that the distance from spiracle to bulla relative to pronotum width differed significantly among the minor workers of host and parasite (F1298 = 551.36, P <0.01). Two groups are clearly shown in Figure 15: one had a small number (n = 25) of parasite workers and another formed by a large number of host minor workers (n = 275). The fact that the workers sorted into two groups, as well as morphological differences between the groups, is highly suggestive. We found some A. ameliae workers are larger than host minors but this is because larger host minors were not sampled. Preliminary genetic analysis by RAPD (Random amplified polymorphic DNA) markers clearly shows differences between the two groups, confirming these results. As in A. insinuator , the workers of A. ameliae have a significantly smaller distance from their spiracle to bulla than their host minor workers of same pronotum width (Figs. 16 and 17). These results are almost identical to those obtained by Sumner et al. (2003) for A. insinuator .
Comments on biology There are previous reports on polygyny in the two host subspecies (Delia Lucia & Vilela, 1989; Delabie, 1989). During nest collection, we indeed found three monoginic parasitized nests which had 2, 3 and 4 queens of A. ameliae . The first collection of the parasite was in October, 2003, when two nests were collected with hundreds of alate males and queens of the parasite. In April of the following year, we again collected nests with the alate parasites. This suggests that the production of the reproductive caste in A. ameliae may occur throughout the year. In the laboratory, males and queens flew towards the light, indicating that this species is likely to perform the nuptial flight in nature. As in A. insinuator , Acromyrmex ameliae produce a workforce. This seems to be essential for the production of the parasite alates (Sumner et al., 2003), but this trait is being selected against over evolutionary time, although it has not yet been lost. We need to investigate if the host cares for the parasite alates. In this case, parasite workers may not be needed.
Paratypes 20 queens, 20 males and 18 workers. Labeled 'Brazil: Paraopeba MG/ 20 April 2004/ I. M. F. Soares (MZUSP).
Discussion
Acromyrmex ameliae is a social parasite with much smaller reproductives (females and males) than those of its hosts (Table 1). Morphometrically, the A. ameliae queen is not a simple miniature of its hosts' queens, like Myrmica microrubra and its host Myrmica rubra (Steiner et al., 2005). Here, we can distinguish the new species from the other of the group with propodeal spines: they are straight and laterally compressed unlike subterraneus subspecies where they are slight to strongly curved and conical. Acromyrmex ameliae differs from A. insinuator not only by its size (compare Tables 1 and 2) and color (brown dark against yellowish-orange) but as well it does not present a single strong median ruga extending from the central ocellus to the level of the posterior borders of lateral ocelli, like A. insinuator . On the contrary, around its central ocellus, the cuticle is wholly rugous without a distinct median ruga. In A. insinuator the anteroventral edge of the postpetiole is broadly and evenly concave, without a broad median anteroventral extension. The anteroventral portion of the post-petiole in A. ameliae has irregular extensions, without the concavity present in the first species.
HL (mm)HW (mm)WL (mm)QueensA. echinatior2.2 ± 0.12.6 ± 0.23.8 ± 0.1(n = 11, 9 nests)A. insinuator1.9 ± 0.12.3 ± 0.13.3 ± 0.2(n = 25, 5 nests)MalesA. echinatior1.4 ± 0.11.8 ± 0.13.5 ± 0.3(n = 12, 11 nests)A. insinuator1.4 ± 0.11.6 ± 0.13.0 ± 0.2(n = 18, 5 nests)
As in Acromyrmex insinuator , reproductives (females and males) A. ameliae very much resemble the host species, although there has been a pronounced reduction in body size. From observations of nuptial flights that occurred in the laboratory, we can suspect that a mating flight occurs in the field. However, observations of nuptial flights in the field are necessary to conclude this. We found alate parasites in two different seasons (April and October), unlike the host species, which has only a single synchronized nuptial flight per year in November and December. More than one nuptial flight each year could increase their likelihood of successful invasion of new colonies. The well defined nuptial flight of the hosts is normally observed in November and December so that newly fertilized parasite queens (produced in April) can colonize established colonies of A. subterraneus subterraneus and A. subterraneus brunneus well before they reproduce themselves. We suggest that further studies should investigate the chemical profile of the involved species to answer the following questions. Does the newly fertilized parasite queen lack external chemical substances, as hypothesized by Lenoir et al. (1999)? Is there an absence of colony specificity in hydrocarbons composition among hosts and parasite species? The chemical ecology of interactions between ants and their social parasites exhibits a potential field for research since it constitutes a multiplicity of complex interactions, yet it has been little studied (Lenoir et al., 2001).
Like A. insinuator (Bekkevold & Boomsma, 2000), A. ameliae is one of the rare inquiline parasites that produces workers and resembles its host in morphology, but unlike A. insinuator , A. ameliae queens differ dramatically from their host queen in size. We excavated 14 colonies of A. subterraneus and found all of them parasitized by A. ameliae . Thus, A. ameliae appears to be very common, yet always overlooked in the past. Perhaps many other similar social parasites await discovery.
