Morphological Description of the Immatures of the Ant, Monomorium floricola

Some ant species of the genus Monomorium Mayr occur worldwide and are considered important urban pests. The larvae of only a few species of this genus have been described, and these descriptions are either superficial or incomplete. This study aimed to determine the number of larval instars and describe the immature stages of the ant Monomorium floricola Jerdon (Formicidae: Myrmicinae). Specimens were analyzed and measured using light and scanning electron microscopy. Three larval instars were found, and all larvae had pheidoloid bodies with ectatommoid mandibles, consistent with other Monomorium species described previously. Five types of body hairs were described, and their distribution was instar-specific. Body and mandible dimensions of the larvae also were constant for each instar. Like other Myrmicinae, the larvae did not create a cocoon. Some of differences among the hair types and sensilla were observed by comparing the samples with larvae of other species in the genus, and these differences may have taxonomic utility.


Introduction
Monomorium Mayr is a cosmopolitan ant genus that includes some 586 described species and subspecies (Bolton et al. 2006), some of which were spread worldwide by commerce and are now regarded as urban pests (Bolton 1987). The most important species include Monomorium destructor Jerdon, Monomorium latinode Mayr, Monomorium floricola Jerdon, Monomorium pharaonis L., Monomorium subopacum Smith and Monomorium talpa Emery (Bolton 1987). Monomorium floricola (Formicidae: Myrmicinae) and M. pharaonis are considered of even greater importance as they may be active carriers of pathogens inside hospitals (Campos-Farinha et al. 2002).
Morphological descriptions of ant larvae can provide new characters that can be useful taxonomic characters, especially for particularly difficult groups of species (Wheeler and Wheeler 1976;Schultz and Meier 1995;Pitts 2005). Moreover, improved morphological knowledge about immature stages of ants may help clarify various aspects of their biology and social organization (Fox et al. 2007).
Most ant larval descriptions were done by George C. and Jeanette Wheeler in a long series of publications describing the larvae of nearly 800 ant species (Wheeler and Wheeler 1988). However, these descriptions were usually based on few specimens and some were made without knowledge of what larval instar was being described. The latter deficiency is a consequence of the difficulties of accurately determining the number of larval instars for most insects. Of more than 11,000 extant ant species (Bolton et al. 2006) the number of larval instars is only known for 64 species from nine subfamilies, and this number typically varies between three and five instars. An updated list of these publications is presented in Table 1.
The number of Monomorium species that had their larvae described (Wheeler and Wheeler 1955, 1960, 1973, 1977Petralia and Vinson 1980;Berndt and Kremer 1986;Wheeler 1989, 1991) was only 2% of the total number of species in the genus. The description of larvae M. floricola was made originally by Wheeler and Wheeler (1955); however, apart from the limitations mentioned above, it lacks important details such as body measurements.
Given the importance of larvae in ant colonies, the fact that there is no information about the number of larval instars for nearly all known ant species is a concern. Only in the last few years have some authors devoted studies to filling in this gap. Such basic knowledge is crucial to understanding immature development inside ant colonies and thus colony cycles and internal dynamics.
This investigation was aimed at determining the number of larval instars of M. floricola workers by measuring the growth rate between each instar, and making a detailed description of every immature stage by light and scanning electron microscopic observations.
The present work focused on describing worker larvae only, as reproductive larvae can be very similar to worker larvae and thus difficult to distinguish. Knowing the general morphology of worker larvae will help identify reproductive larvae. Reproductives are usually produced during intermittent periods in much lower numbers. Therefore, care was taken to collect larvae only during periods when no reproductive forms were being produced. Workers of this species are  Wheeler and Wheeler (1976).

Determining the number of larval instars
As we were unable to directly observe moults, the number of larval instars was determined using the method described in Parra and Haddad (1989). The maximum head widths of the larvae were measured (n = 344) and plotted in a frequency distribution graph, wherein every distinct peak was considered to correspond to a different larval instar; the obtained number of larval instars was then tested against Dyar's rule (Parra and Haddad 1989). The first larval instar and the last larval instar can be explicitly identified and used as reference to bracket others. First instar larvae are equivalent to the mature embryo, which can be measured in the egg through the transparent chorion, and last instar larvae have the developing pupa showing from within (also termed 'prepupae').

Description of the immature forms
The morphological observations were made with light microscope (Zeiss MC80 DX, with maximum magnification of 1000X, www.zeiss.com) and a scanning electron microscope (Phillips SEM-505, at 12.0 kV), using ten larvae of each instar for each method. The ten larvae were selected among the ones presenting the most frequent head width found for the respective instar. With a stereomicroscope (Zeiss Stemi SV11, with maximum magnification of 66X) equipped with a micrometric eyepiece, the length and width of eggs (n = 159) and larvae (n = 50 for each instar), and length of pupae (n = 50) were rapidly measured. Also, body length between spiracles, a mode of measuring larvae devised by Wheeler and Wheeler (1976) that accounts for body curvature, was determined for 10 larvae of each instar. All measures are given as mean ± standard deviation, where applicable.
All collected samples were fixed in Dietrich's solution (900 ml distilled water, 450 ml 95% ethanol, 150 ml 40% formaldehyde, 30 ml acetic acid) for 24h. The samples were then transferred and conserved in 80% alcohol. For scanning electron microscope analysis, the samples were dehydrated in an acetone graded series (70-100%; specimens dipped for five min in each concentration), and critical-point dried (Balzers CPD/030, www.balzers.com). Dried specimens were then attached to aluminium stubs with double-faced conductive, adhesive tape and were goldsputtered with a Balzers SCD/050 sputterer. Observations and images were obtained as soon as possible from sample preparation. Prior to analysis under the light microscope, the larvae were warmed for 10 min in KOH 10% and placed in a small drop of glycerine on a microscope slide.

Determination of number of larval instars
The frequency distribution of the maximum head widths of the larvae formed a multimodal curve with three distinct peaks (Figure 1), suggesting the existence of three larval instars. The number of instars proposed by our results yielded a good fit with Dyar's rule (R = 0.95). Mean growth rate through all instars was 1.23, while the growth rate was 1.23 between the first and second instars, and 1.22 between the second and third instars. Descriptions of the immature stages are as follows.
HEAD CAPSULE (Figure 3d): Cranium 0.15 mm wide (n = 38); subcircular. Antennae with three basiconic sensilla over a slight elliptical elevation (0.006 mm wide and 0.004 mm high) (Figure 3e). Twenty-six head hairs of two types: 4 type D hairs on the ventral border of the clypeus; 6 hairs over gena, 4 of type C and 2 of type D (the latter placed near the border of the clypeus); 2 type C and 2 type D hairs over frons (the latter over the frons bordering the clypeus); 2 type C hairs over vertex; and 4 type C hairs along occipital border. Type C hairs 0.022 -0.030 mm long (n = 4) and type D hairs 0.015 -0.023 mm long (n = 9). Well defined clypeus with no sensilla.

Pupa
Exarate, with no cocoon; whitish when young, with surface and eyes getting darker on late metamorphosis. Body length 1.42 ± 0.07 mm, varying 1.26 -1.54 mm (only white pupae were measured).

Determination of number of larval instars
The number of instars herein recorded for M. floricola was the same as in M. pharaonis (Table 1). The occurrence of three larval instars is known to four ant subfamilies. This was recorded to 13 myrmicine species including M. pharaonis (Table 1).
The larval growth rate observed agrees with the Dyar principle, which states that the head capsule of larvae grows in geometric progression over the ecdises at a constant rate that varies between 1.1 and 1.9, usually around 1.4 (Parra and Haddad 1989).

Morphological description of the immature forms
The original larval description of M. floricola by Wheeler and Wheeler (1955) was based on their description in the same paper of M. pharaonis. Therein the larvae were separated as 'very young' and 'young' specimens, possibly corresponding to the first and third instars judging from the stated specimen size and body hairs. Their number of analyzed specimens was specified as "numerous." Moreover, in that description, many body measures of the larvae were not taken. In the present study, many gaps have been filled in by measuring mouthparts, head capsule, spiracles, body width and length, and the body length between spiracles. Wheeler and Wheeler (1960) attempted to set parameters for a genus-level taxonomy of ant larvae. The main traits they considered when building an identification key to ant larvae was the body shape of the larvae in side view, the shape of mandibles and the hair types present. These traits are discussed, in this same sequence, below, while comparing them in different related ant species.
The body shape herein observed for M. floricola larvae agrees with observations by Wheeler and Wheeler (1976) for Monomorium.
From comparing the body measures of M. floricola larvae with those of M. pharaonis (Berndt and Eichler 1987), it was verified that both species have first instar larvae of approximately the same size, while second and third instar larvae of M. pharaonis are always larger (first instars measure 0.39 mm long and 0.18 mm wide, while the others respectively measure 0.60 mm long and 0.26 mm wide, and 1.27 mm long and 0.59 mm wide).
The mandibles of M. floricola larvae confirm the statement made by Wheeler and Wheeler (1976) that all hitherto described Monomorium larvae have ectatommoid mandibles, thus this trait remains reliable for separating larvae from different genera of Myrmicinae. Based on the observations outlined here, one can promptly sort out first instar larvae of this species as those with completely unsclerotised mandibles. This is useful because directly dissecting and measuring mandibles is very difficult.
In addition to two hair types previously recorded by Wheeler and Wheeler (1955) for M. floricola, three other types have been identified. Body hair distribution and types of hairs was typical of different instars (Table 2): simple body hairs were typical of first instars, while the curved bifid, multifid and unbranched denticulate hairs were typical of second and third instar larvae. Denticulate bifid hairs were found in all. All these hair types, except the unbranched denticulate ones, were also observed in larvae of M. pharaonis (Wheeler and Wheeler 1955;Petralia and Vinson 1980;Berndt and Eichler 1987), with the latter also having two different hair types: deeply three-branched with tips curved in opposite directions (Wheeler and Wheeler 1955) and another unbranched hook-like smooth hair (Berndt and Eichler 1987). Monomorium antarcticum Smith larvae had three hair types (Wheeler and Wheeler 1955), including one similar to the unbranched denticulate hairs found in M. floricola occurring on the thoracic area. Monomorium afrum André larvae were reported to bear two types of hair (Wheeler and Wheeler 1955), with one of them similar to the bifid curved hairs herein described.
All the head hairs of the mature larvae observed in this study were branched, in accordance with the Wheeler and Wheeler (1960) identification key for Myrmicinae larvae, with the exception of M. antarcticum. The use of head hairs differences between larvae of different species was proposed by Pitts et al. (2005) to aid species identification in the troublesome Solenopsis saevissima Smith complex of fire ant species. Given the proximity of Solenopsis to Monomorium (Bolton 1987), it is possible that a similar approach would be useful in separating larvae of closely related species. It should be noted that Fox et al. (2007) detected intraspecific variation in head hair types in Paratrechina longicornis Latreille larvae of the same instar and questioned the true applicability of such characters, while also remarking on the low number of specimens analyzed for Wheelers' descriptions. Moreover, in the present description variation in the morphology of head hairs was detected among different specimens of first instar larvae.
In comparing the labrum of mature larvae of M. floricola with those of the other species described in Wheeler and Wheeler (1955), it was observed they differ on the labrum from M. pharaonis by having between six and eight bifid hairs on the anterior surface, where mature larvae of M. floricola have six simple hairs. Larvae of M. antarcticum present between ten and twelve short hairs or sensilla on the labrum (Wheeler and Wheeler 1955).
The sensilla on the ventral surface of the labrum, as well as the sensilla near the extremities of the spinneret, base of mandibles and on the antennae, were not observable in first instar larvae, although some of them might be present. These specimens were covered in debris precluding the observation of details. The second and third instar larvae possessed hairs on the maxillae and setaceous sensilla on the maxillary palps and galea, while first instar larvae had only basiconic sensilla and no hairs on the maxillary region and parts.
Ant larvae generally have ten pairs of spiracles, two thoracic and eight abdominal; in 50% of the genera, spiracles are relatively small and uniformly sized (Wheeler and Wheeler 1976). The first thoracic pair of spiracles of M. floricola was always bigger than the others, which are all uniform. The same was observed in M. pharaonis, Monomorium tambourinensis Forel, and M. antarcticum (Wheeler and Wheeler 1955;Wheeler and Wheeler 1960); this could be a recurrent trait in this genus. This was already proposed by Wheeler and Wheeler (1976), but other species must be looked at in order to confirm the pattern.
As stated by Wheeler and Wheeler (1976), all ants have silk glands, but some of them, including the Myrmicinae, do not weave cocoons. M. floricola fit this pattern.
Finally, this study presented a deeper analysis of the morphology of the larvae of M. floricola, previously described by Wheeler and Wheeler (1955), while comparing it with the few other Monomorium larvae described so far and confirming general similarities, e.g. the body and mandible formats. Also, the number of instars for this species was established. The different types of hair described and other findings that have direct application for instar separation with this species are particularly important. Some of the information herein may aid future systematic and taxonomic studies with the group, as well as help clarify some aspects of the biological and social organization of this ant.