Overview of the Morphology of Neotropical Termite Workers : History and Practice

This contribution deals with the worker caste of the Neotropical termite fauna. It is a compilation of present knowledge about the morphology of pseudergates and workers, including the literature discussing the origin and evolution of this caste, the terminology used in the different taxonomic groups, and the techniques used to study these individuals, especially examination of the gut, mandibles, legs, and nota. In order to assist in identifying workers, it includes a key for the families that occur in the Neotropical Region and a characterization of workers of all families, especially the subfamilies of Termitidae, with descriptions and illustrations of diagnostic morphological features of genera. We point out advances and gaps in knowledge, as well as directions for future research.

Overview of the Morphology of Neotropical Termite Workers: History and Practice

Introduction
Adapted to a strict diet of cellulose and lignocellulose compounds, termites are eusocial hemimetabolous insects that live in organized colonies, with castes that can be recognized morphologically and functionally.Although the group contains only a moderate number of species for insects (approximately 3000 species worldwide and 650 Neotropical species), and they are most commonly known as pests (structural, agricultural, and forest), termites have critical roles in tropical and subtropical ecosystems, and are hence termed ecosystem possibilities are many, complex and widely debated (e.g., Roisin & Korb, 2011).Thus, the termitology jargon includes several terms that merit more-detailed explanation (see Thorne, 1996;Korb & Harfelder, 2008).Here we will treat the terms relating to juveniles, and subsequently those for workers.
As the soldier caste shows distinct characters, termites have traditionally been identified based on this caste.However, there are practical reasons for attempting to use the worker caste for identification.These include the identification of specimens in samples that lack soldiers (such samples are commonly left unidentified in collections); triage of miscellanea provided from field work; and to better understand the relationships among taxa, i.e., to understand the evolution of the group, since worker characters provide substantial taxonomic information (Donovan et al., 2000;Inward et al., 2007;Rocha et al., 2017).
A fine history of termitology was provided by Krishna et al. (2013), who highlighted the role of Niels F. Holmgren as the most important researcher to study the morphology of termites.Holmgren (1909) described the gut in situ of imagoes, soldiers, and workers of the same species in several genera representing different groups.Holmgren's work was essential for a classification system of the group, and laid the foundation for all who came after him.Some early authors investigated the digestive tract of termites, although they were concerned with studying the intestinal symbionts, initially called parasites (e.g., Lespès, 1856, Leidy, 1881, and Grassi & Sandias, 1893).Cleveland (1924), in his seminal work on the relationship between protozoan symbionts and their hosts, provided a detailed history of the studies that dealt with these symbionts, emphasizing that initially, the authors studied either the termites or the protozoa, but few attempted to understand the relationship between them.He cited Buscalione and Comes (1910) and Imms (1920), who referred to a symbiotic relationship between the intestinal protozoa and their hosts, but without a solid experimental basis.Through hundreds of laboratory experiments, Cleveland himself definitively proved that a true symbiotic relationship exists between some of the intestinal protozoa and the termites, and even demonstrated that some other intestinal protozoa were not symbionts but merely commensals.He also stated that no member of Termitidae contains protozoan symbionts, which was evidenced by many later studies.Today there is a vast literature on the intestinal symbionts of Termitidae, which include bacteria and fungi (König & Varma, 2010;Ohkuma & Brune, 2011;Brune & Dietrich, 2015, among others).
Pierre Paul Grassé and Charles Noirot studied many African species, also investigating the digestive tract of workers.In 1955, they proposed a new subfamily, Apicotermitinae (Termitidae), based mainly on the gut anatomy of the workers.They provided detailed comparisons between the new subfamily and the very concept of Termitinae at that time, for both soldiers and workers.
Even after this study, few taxonomic descriptions at that time included such characteristics (Noirot, 1966;Deligne & Pasteels, 1969).Noirot and Kovoor (1958) contributed significantly to knowledge of the comparative anatomy of the gut of African Termitinae, from the perspective of understanding the evolution of these characters.Jacqueline Kovoor (1969) continued these studies of comparative anatomy, including photographs of the enteric valve for the first time.She studied the Nasutitermitinae, which at that time included the genera which are today in Syntermitinae, that is, she also studied Neotropical species.This study was definitive for all researchers who continued to work with termites in the Neotropical Region.
Noirot and Noirot-Thimothée (1969) composed the chapter on the termite digestive system in the first volume of a two-volume work, the fundamental Biology of Termites (Krishna & Weesner, 1969, 1970), to this day an obligatory reference for termitologists and a "Bible" for the Isoptera until the publication of the treatise "Termitologia" by Grassé (1982Grassé ( , 1984Grassé ( , 1986)).Sands (1972), in his study on soldierless termites of Africa, proposed a new classification of the Termitidae, broadening the definitions of Apicotermitinae (to include Anoplotermes) and Termitinae (to include Amitermitinae).For this purpose, he studied in detail the general external morphology, morphology of the mandibles, and the anatomy of the gut of workers.Johnson (1979) perceived that the greatest problem in termite population ecology was to identify workers, which were often collected from soil as mixtures of species.He developed the convention, still in use today, for illustrating the in situ gut morphology of Termitidae workers (e.g., Figs 1-4).While Johnson's (1979) studies involved African taxa, Mathews (1977), in his study on the termites of the Mato Grosso region of Brazil, recorded some characters of the gut of certain species (enteric valve and/or uncoiled gut of new genera of Apicotermitinae).
From the 1980s onward, termitologists began to describe workers more completely, especially with descriptions of the gut (Miller, 1984;Fontes, 1985;Constantino, 1991bConstantino, , 1995;;Roisin et al., 1996, among others).Now in the 21st century, virtually all descriptions of new Termitidae taxa and revisions include a study of the gut and mandibles of the workers (e.g., Cancello & Myles, 2000;Cancello & Noirot, 2003;Bourguignon et al., 2008Bourguignon et al., , 2010;;Carrijo et al., 2011;Cuezzo & Nickle, 2011;Constantino & Carvalho, 2012;Rocha et al., 2012;Cuezzo et al., 2017;Hellemans et al., 2017).Noirot (1995Noirot ( , 2001) compiled all the studies on the species of all families of Isoptera, with illustrations and descriptions, including important discussions about the evolutionary significance of the differences among the various parts of the gut, especially in the Termitidae (Noirot, 2001).These two studies are a fundamental basis for this area of research.Even though it deals with African fauna, one cannot fail to mention the book by Sands (1998) on the workers of African termites, with its 1568 drawings and 18 photographic plates.
The descriptions of the external morphology and of the mandibles are exemplary, and those of the gut are definitive.For a better understanding of the worker caste, it is also important to mention studies of the mandibles, such as those of Ahmad (1950), who proposed a phylogeny of the Isoptera (today completely outdated) based only on the imago-worker mandibles; and Deligne (1966Deligne ( , 1999)), who dealt in detail with the functional morphology of the mandibles and their modifications during development.
The origin and evolution of the worker caste have been a subject of discussion, and several studies have investigated this fundamental aspect of termite colonies in the different taxonomic groups (Noirot & Pasteels, 1987;Korb, 2007Korb, , 2008;;Inward et al., 2007;Legendre et al., 2008;Roisin & Korb, 2011;Bourguignon et al., 2017).In recent years, remarkable insights have been gained on the termite worker caste, with important implications, for example, for the genetic, hormonal and ecological influences on caste differentiation, developmental pathways, and polyethism, and the microbial symbionts in the gut and their co-evolution with termite hosts.This knowledge reinforces the importance of the role of the worker caste and the successful evolution of the group (Bignell et al., 2011).
This review summarizes recent progress in morphological characterization of workers of the Neotropical termite fauna, covering all families and subfamilies that occur in this region and their importance in taxonomy and phylogeny.Much advancement is needed, but we indicate several further steps that may accelerate taxonomic decisions and help in other areas such as structural, agricultural, and forest "pest species".We hope that with the present study and those to come, knowledge and identification of workers will become more and more common in the day-to-day routine of the termitologist.
Since termites are hemimetabolous insects, all juveniles should be called nymphs, but the term "larva" has become classic for the first-instar juveniles, with no wing bud or soldier features; and the term "nymph" is used for juveniles with wing buds of different sizes, that is, which are in the imaginal lineage (Roisin & Korb, 2011).The soldier before the last molt is called "white soldier" and already exhibits some of the distinct features of the soldier of the species.The so-called "white-gutted soldier (WGS)" (Scheffrahn et al., 2017) in contrast, is fully developed and has functional mandibles Most simply, one can define the different types of workers as follows (adapted from Korb & Hartfelder, 2008): 1) Pseudergate sensu stricto: in all families except Termitidae, the "larvae" are totipotent with respect to development, and can become soldiers or reproductives, or can undergo "stationary molts", without developing wing buds or characteristics of soldiers.They can even undergo regressive molts, when the nymph with a wing bud molts to an instar without a wing bud.Grassé and Noirot (1947) coined the term "pseudergate" for a juvenile that undergoes a regressive molt.However, this raises practical problems, because it is necessary to follow the development of a nymph with wing buds to discover whether it goes through a regressive molt and loses them.2) Pseudergate sensu lato: In this broader definition, any "larva" or nymph that could potentially undergo stationary, regressive or progressive molts is called a "pseudergate".
In both the above definitions, the individuals are little sclerotized and whitish, and may be functional workers.Other terms exist, such as "false worker" and "helper", with different definitions, including in relation to the work that they perform in the colony (see discussions by Noirot & Pasteels, 1987).Here, "pseudergate" is used in the broader sense.
3) True workers: These are individuals that differentiate early and irreversibly from the imaginal line of development, and are less flexible than "pseudergates", i.e., they are not totipotent and cannot undergo regressive or stationary molts.Morphologically they lack wing buds and have rudimentary sexual organs.These individuals are responsible for practically all the tasks of the colony except reproduction, including defense.True workers are present in Mastotermitidae, Hodotermitidae, in some Rhinotermitidae, and in all Termitidae.Here, the term "worker" is used.

Digital images and drawings
The line drawings were made with a camera lucida, and the photographs were taken with a digital camera coupled to a stereomicroscope, at different focal points, and merged with software.The enteric valve and crop of workers were mounted on a slide with glycerin and photographed under an optical microscope.
Readers should note that in older workers the mandibles often appear worn, with the tips of the teeth not pointed, tapered but blunt; and the edges of the marginal teeth may appear to be interrupted or serrated due to wear.
Another factor is the degree of filling of the gut, which may alter the relative position of the parts from that in published illustrations and photographs.To prevent large differences between descriptions, it is useful to inspect several individuals before selecting one for illustration.Normally, the relative position does not change so much that it prevents the recognition of the patterns illustrated in more recent studies.Porotermes has three extant damp-wood species: P. adamsoni, native to coastal and adjacent highland areas from southern Queensland to South Australia, and considered an invasive termite in New Zealand (Pearson et al., 2010); P. planiceps, from the western Cape region of South Africa (Coaton & Sheasby, 1976); and Porotermes quadricollis, in southern Chile and Argentina (Camousseight & Vera, 2005;Torales et al., 2005).Villan (1972) redescribed Porotermes, with new diagnoses and a key to the three species, for imagoes, soldiers, and pseudergates.

Key to the families found in the
Characteristics of the pseudergates of Porotermes are: compound eyes more or less developed, depending on the stage of development of the individual; ocelli absent; pronotum half-moon-shaped; tarsi with three tarsomeres on each leg, with a triangular arrangement on the posterior leg; cerci with five articles, styli are present in males and females.
For identification of the pseudergates, in addition to the mandibular pattern, Noirot (1995) described the gut coiling in situ and gave details of its anatomy in P. adamsoni and Stolotermes africanus.Krishna et al. (2013) described the gut of Porotermitinae in detail and compared some characters with other families.
Compared to non-wood-nesters, the pseudergates of both Kalotermitidae and Stolotermitidae show proportionally short legs, with an enlarged coxa; and their behavior is different, as they live inside wood, crawling in very slow movements.

Kalotermitidae
Kalotermitidae is represented worldwide by 22 genera and 459 extant species.The Neotropical Region has the largest number of genera (12), although the Oriental Region has the most species diversity.The so-called dry-wood termites, e.g., certain species of Cryptotermes, and Incisitermes, can cause enormous damage in urban areas.Constantino (2002) mentioned three species of this family that were introduced from other regions and became urban pests: Cryptotermes brevis, C. dudleyi, and C. havilandi, and noted that there are several yet-undescribed species of Cryptotermes in South America.In Brazil, C. brevis is a major pest in many cities (Constantino, 2002).We may expect that some species, originally restricted to the natural environment and coming into contact with cities expanding into pristine forests, may become urban pests.An example is Glyptotermes canellae, an Atlantic Forest species reported from inside wooden furniture in Santos, state of São Paulo (E.M.C. personal observation).
True workers are absent, and pseudergates contribute to the tasks within the colony.In general, given the uniformity of their external morphology, only the mandibles are usually used for identification, even though they show a similar pattern, as seen in the illustrations by Krishna (1961).No specific diagnoses were provided for pseudergates until Scheffrahn (2011) described the pseudergates of Calcaritermes, illustrating their mesonotal "rasp" and/or the concave posterior margin of the pronotum.The mesonotal rasp is the first external character used to provide a diagnostic, generic identification of an immature kalotermitid.

Serritermitidae
Serritermitidae is the only termite family with exclusively extant Neotropical representatives (records from Brazil, Guyana, and French Guiana).The taxon was originally proposed by Holmgren (1910) as a subfamily in the Mesotermitidae (nomen nudum, related to Rhinotermitidae) and raised to family rank by Emerson (1965).The family presently contains three species in two genera, Glossotermes and Serritermes, according to Cancello and DeSouza (2005).
The mandibles of the imago and the pseudergate are unique (Ahmad, 1950;Cancello & DeSouza, 2005), with long acuminate apical teeth, a reduced number of marginal teeth (left mandible with one marginal tooth in Serritermes, and two in Glossotermes), and the molar regions with well-developed ridges (Fig 9).The gut shares similarities with members of the family Rhinotermitidae, except in the reduction of the stomodeal valve and the simplification of the structures present on P1 and P2 (Noirot, 1995;Cancello & DeSouza, 2005).The clear differences between the worker mandibles in the two genera, Glossotermes and Serritermes, are easily seen and are enough to separate them.

Rhinotermitidae
Rhinotermitidae is the most widely distributed termite family in the world, occurring from tropical to temperate regions.Some species of this family, popularly known as subterranean termites, are especially important because of their potential as pests, and can cause enormous damage (Constantino, 2002;Vargo & Husseneder, 2009).
Rhinotermitidae was initially described as a subfamily of Mesotermitidae (nomen nudum, Holmgren, 1910), and was later raised to family rank.The relationship between the genera of Rhinotermitidae and the other termite families, especially Termitidae, Serritermitidae, and Stylotermitidae, has been widely debated.The present consensus is that Rhinotermitidae is not a natural group (Lo et al., 2004;Inward et al., 2007;Legendre et al., 2008Legendre et al., , 2013;;Bourguignon et al., 2014).The family is composed of the genera Prorhinotermes (three species in the Neotropical Region), Termitogeton, and Psammotermes (outside the Neotropical Region), of uncertain position, although each was classified in a different subfamily by Krishna et al. (2013); and two relatively consistent groups: the Rhinotermitinae, composed of Parrhinotermes, Schedorhinotermes, and Macrorhinotermes (outside the Neotropical Region); plus Acorhinotermes, Rhinotermes, and Dolichorhinotermes (Neotropical); and the group composed of Coptotermes, Heterotermes, and Reticulitermes (Bourguignon & Roisin, 2011;Bourguignon et al., 2014).This last group was traditionally divided into two subfamilies, Coptotermitinae (Coptotermes) and Heterotermitinae (Heterotermes and Reticulitermes); however, in all recent phylogenies using molecular data the genus Coptotermes supports the subfamily Heterotermitinae as paraphyletic (e.g., Lo et al., 2004;Inward et al., 2007;Legendre et al., 2008;Bourguignon et al., 2014).Thus, we use here "Heterotermes group" for the three genera Coptotermes, Heterotermes, and Reticulitermes.The Heterotermes group is probably a sister-group of the Termitidae and will likely be raised to family rank (Lo et al., 2004;Legendre et al., 2013;Bourguignon et al., 2014).
Workers of Rhinotermitidae can be easily recognized by the three marginal teeth on the left mandible, and a subsidiary tooth on the right mandible (Fig 7-8).The gizzard is shorter and considerably more separated from the esophagus than in the other, non-Termitidae families.The transition from P1 to P2 is well marked.P1 is short and narrowed distally.The transition from P1 to P2 has six cushions of two orders (I and II).In P1, the cushions have small homogeneous spines, with cushions I nearly rectangular, and II triangular and narrowed distally.Cushions I are well developed and ornamented in P2; Noirot (1995) described them as a "sharp-pointed escutcheons", with the distal tip protruding at the opening of the P2 in P3.Cushions II are reduced in P2.
The ornamentation of the "escutcheons" is somewhat characteristic of the genera and/or groups of genera.In Prorhinotermes, the "escutcheons" are elongated and arranged in a proximal cluster with 3-4 rows of one-, two-, or three-pointed scales (Fig 20).In the Rhinotermitinae, the proximal cluster has scales with larger spines in the center of the "escutcheons", but not as large as the spines in the Heterotermes group (Figs 18,19).Both Rhinotermitinae (Fig 21) and the Heterotermes group also have two distal clusters of fine setae, and Rhinotermitinae has a central cluster of spines that are less sclerotized than the proximal cluster, but more so than the corresponding setae of the distal clusters (Noirot, 1995).
Among the groups with Neotropical representatives, the external morphology of the workers is quite varied.In workers of Rhinotermitinae, the second marginal tooth of the mandible is smaller than the first marginal tooth (Krishna et al., 2013); the pronotum is saddle-shaped; the wall of the abdomen may be transparent or slightly opaque; and the head is large in relation to the body size (Fig 17).In workers of the Heterotermes group, the first marginal tooth of the left mandible is smaller than (Coptotermes and Heterotermes) or equal to (Reticulitermes) the apical, second and third marginal teeth (Krishna et al., 2013); the pronotum is not saddle-shaped as in Termitidae, with at most a small elevation on the anterior margin (i.e., a very small anterior lobe); the head is relatively smaller in relation to the body; and the abdomen is always whitish and opaque (Fig 16).

Termitidae
This family has the highest diversity of extant genera (261) and species (2097), with 81 genera and 428 species in the Neotropical Region.They occur in a wide diversity of habitats, mainly in tropical and subtropical regions worldwide.
Species of Termitidae exploit the most varied types of lignocellulosic substrates as food.Most of the genera of Nasutitermitinae can be considered xylophagous (e.g., Cortaritermes, Caribitermes, Nasutitermes); others are humivores (e.g., Cyranotermes, Subulitermes); some feed on leaf litter (e.g., Velocitermes); or have specialized diets, such as some species of Constrictotermes that have peculiar feeding habits such as scraping lichens and tree bark.The diets of Syntermitinae encompass the entire humification gradient (Rocha et al., 2017), so that among its species it is possible to find all the morphotypes associated with their respective food resources.Among the Termitinae, again, the species cover almost the entire humification gradient; some are strict xylophages (e.g., Cylindrotermes) and others are humivores (e.g., Cavitermes) or are intermediate.Most members of Apicotermitinae are restricted to a humus-rich diet, and others, such as several species of Ruptitermes, eat grasses.
A worker caste can be recognized in the Termitidae, with a line of development, morphology, and distinct functions within the colony.Importantly, the mandibles of Termitidae are not the only components needed for rapid identification, as in the other families, but must be considered together with the characters of the gut and external morphology for an accurate determination.
The morphological characters that allow identification of workers have been treated in several studies over time.For example, the numbers of antennomeres and tibial spurs, in addition to the mandibles, were reported in detail by Weesner (1969), Grassé (1982), and Krishna et al. (2013).Characters of the internal morphology such as the gut coiling in situ, the mixed segment, the shape of the mesenteric tongue, and the insertion and arrangement of the Malpighian tubules (usually numbering four), and the enteric valve are derived mainly from the studies by Grassé and Noirot (1954), Noirot and Kovoor (1958), Kovoor (1969), Sands (1972, 1998), and Noirot (2001).

Apicotermitinae
A total of 50 genera and 217 species are recognized worldwide, with 14 genera and 52 species in the Neotropical Region.Advancement of taxonomic knowledge of Apicotermitinae in the world, and particularly in the Neotropical Region, presents a difficult challenge.Partly this is due to the absence of soldiers in most of the genera, and researchers have concentrated, consequently, on the identification and delimitation of the taxa based on the characters of the gut, which necessarily requires more time and training.Another consideration is the high rates of abundance and diversity of Apicotermitinae in the various ecosystem types, especially in humid forests, emphasizing their ecological importance (Cancello, 1994;Eggleton et al., 1996;Kambhampati & Eggleton, 2000;Davies et al., 2003;Ackerman et al., 2009;Bourguignon et al., 2011;Palin et al., 2011;Cancello et al., 2014).
The group is presently under study in different laboratories, which have a growing interest in improving taxonomic knowledge of the group and advancing studies in other areas (phylogeny, biogeography, phylogeography, ecology, and behavior).
For 104 years, the genus Anoplotermes united all the Neotropical species of termites without soldiers, until the descriptions of Grigiotermes and Ruptitermes by Mathews (1977) and of Aparatermes and Tetimatermes by Fontes (1986).These five genera remain the references for non-taxonomist termitologists who use the most widely disseminated keys for the group.
The available keys for workers of Apicotermitinae are those of Fontes (1985Fontes ( , 1992)), based mainly on characters of the gut coiling; Constantino (1999Constantino ( , 2002)), who concentrated on characters of the external morphology; and Bourguignon et al. (2016), who used the ornamentation of the enteric valve.This last is the most up-to-date key at present, although it does not include Disjunctitermes or Echinotermes.These keys can serve as a point of departure for identification, but as discussed above, the great diversity among the Apicotermitinae remains understudied.Attempts to group new taxa within these known genera have been very common in surveys of the termitofauna, which are consequently inadequate and counterproductive, rendering unreliable those ecological analyses that include the Apicotermitinae.
The Neotropical genera present the following set of characters: the absence of the soldier caste; the presence of an incision between the M1+2 and the M3 on the left mandible (Fig 10,arrow); the P1 long, forming a loop that in ventral view surrounds the P3 (Figs 22,23); and the presence, in most genera, of an enteric-valve seating (EVS, not exclusive to Apicotermitinae) (Noirot, 2001;Fig 24, arrow).The EVS is the distinct proximal region of the P3, shaped as a tubular ring or trilobate, and separated from the main portion of the P3 by a constriction.The P2, with variable organization and ornamentation, is invaginated in the EVS.Certain recurring features aid in the delimitation of morphospecies and establishment of identity for Apicotermitinae; many of these are derived from the studies of Sands (1972Sands ( , 1998) ) and Noirot (2001).The foreleg shows several characters that deserve attention.The foretibia may show a gradient of intumescences, from heavily inflated (Amplucrutermes inflatus, fig 2B of Bourguignon et al., 2016)  In studying the gut coiling, for most groups there is no need for dissection is not needed, since the gut can be examined through the transparent body.In this examination it is important to evaluate certain features that are briefly discussed here.
The mixed segment is an informative region for the study of Apicotermitinae.Noirot (2001)    Among the important characteristics is the enteric valve or P2, which, differently from the mixed segment, requires dissection.Analysis of the P2 under a microscope offers the possibility of studying the arrangement and ornamentation of the scales that form the cushions.As observed for the species of Anoplotermes (including those recently revised by Bourguignon, 2010), the external morphology of the worker is very homogeneous, and it is impossible to distinguish them without examining the P2.The enteric valve is also very important for identification of the species of Patawatermes, although P. turricula and P. nigripunctatus are more easily distinguished through their morphometry and pilosity.
Examination of the enteric valve alone, however, is sometimes not enough to distinguish some Apicotermitinae, such as species of Hydrecotermes (see Bourguignon et al., 2016,

Nasutitermitinae
Nasutitermitinae constitutes a well-established monophyletic group, but the evolutionary relationships among its genera remain unknown (Bourguignon et al., 2017;Inward et al., 2007).The subfamily includes 81 genera and more than 600 species worldwide; of these, 172 species and 31 genera are Neotropical (Table 1).In contrast to Apicotermitinae, Syntermitinae, and Termitinae, in Nasutitermitinae it is possible to recognize more than one type of worker by a mandibular dimorphism (Fontes, 1981(Fontes, , 1987a)).The gap between M3 and the molar plate on the left mandible can be narrow or broad (Figs 37,38).This character has been recorded for all Neotropical genera of Nasutitermitinae except Angularitermes, Cyranotermes, Sinqasapatermes, and Tiunatermes.Considering that this mandibular dimorphism appears not to be associated with sex, it is probably related to the instar of the worker (a narrow gap in the first instar and a broad gap in the second instar), a hypothesis that is presently under investigation.
The characters of the gut are certainly the most important for accurate identification of the workers.The crop of Constrictotermes is strongly dilated and, when full, occupies almost half of the anterior region of the abdomen.The ornamentation of the gizzard is well developed, with both columns and pulvilli; the exception is Agnathotermes, which has the columns completely reduced, leaving only the first-and second-order pulvilli.In turn, the first-and secondorder columns can bear some characteristic structures: The P2 varies widely in its location, and can be found on the dorsal, left or right side of the abdomen.The ornamentation of the enteric valve is widely variable among the genera, as it is among the species, and because of this it deserves special attention in the identification.Generally, the enteric valves of soil-feeding Nasutitermitinae are more variable and complex than those of wood-feeding species.The extent of its variation must be described case by case, which is outside the scope of this contribution, although we mention some references in Table 1.In Diversitermes, Oliveira & Constantino (2016) suggested that an identification of the species based on the soldiers should be confirmed through analysis of the enteric valve of the workers.
Particularly in Araujotermes and Sinqasapatermes the EVS is trilobate.Tenuirostritermes has a diverticulum in the P3.
The distal region of the P3 passes through the mesenteric arch and can be spiral, quite variable in size, or only protrude and connect to the P4 without forming a "dorsal torsion" as in Constrictotermes, Caribitermes, Cortaritermes, Obtusitermes, Parvitermes sensu stricto, and Nasutitermes.In these genera, the isthmus is not apparent in the right dorsal region, which separates the P3 from the P4.In Constrictotermes the distal region of the P3 is completely covered by the crop.The P4 is generally tubular and forms an arch that passes beneath the midgut to the P5.In Angularitermes, Cyranotermes, Ereymatermes, and Sinqasapatermes the P4 is dilated in this passage beneath the midgut.In Antillitermes the P4 is entirely displaced to the ventral region.The P5 is fusiform or globose and has no taxonomic significance.

Syntermitinae
The subfamily Syntermitinae is endemic to the Neotropical Region, and presently includes 18 genera and 103 species (Constantino, 2018).The soldiers have the frontal tube of various sizes, and functional mandibles for defense, which has given them the informal designation of "mandibulate nasutes".The earliest classifications considered the group to be part of the subfamily Nasutitermitinae (Holmgren, 1912;Hare, 1937), which also included the "true nasutes" (whose soldiers have vestigial mandibles).However, phylogenetic studies showed that the frontal tube arose independently in the two groups during the evolution of termites (Inward et al., 2007;  Bourguignon et al., 2017).Even though the "mandibulate nasutes" and the "true nasutes" were recognized very early in the development of termitology, the subfamily Syntermitinae was only formally described by Engel and Krishna (2004), and a more-detailed diagnosis was provided by Constantino and Carvalho (2011).
Among the studies on the gut anatomy of species of Syntermitinae, the pioneer was that of Kovoor (1969), who compared some species of "mandibulate nasutes" and "true nasutes".However, Kovoor (1969) did not question the unity of Nasutitermitinae, and interpreted the differences between the groups in accordance with the proposal of Ahmad (1950), who suggested that the "true nasutes" were derived from the "mandibulate nasutes".
At present, among Neotropical termites the gut anatomy of the Syntermitinae is best known, with consistent and updated descriptions for most of the species and genera (Table 1).There are also studies of comparative anatomy, such as that of Rocha and Constantini (2015), who described patterns of internal ornamentation of the P1 of 36 species of Syntermitinae, including representatives of all the genera.The most recent treatment is the phylogenetic scheme proposed by Rocha et al. (2017), where the authors proposed homologies among gut structures of a representative portion of the Syntermitinae.
When collecting in the field, most samples of Syntermitinae include the soldier caste, mainly when it was collected inside the nests.On the other hand, the nests of Syntermitinae serve as shelter for a large number of other species of Isoptera, of both Syntermitinae as well as other subfamilies, and it is common for samples to contain a mixture of soldiers and workers of many species.Therefore, to correctly screen the material and associate both workers and soldiers with their respective species, knowledge of the worker anatomy is essential.The insertion of the stomodeal valve varies depending on the species and genus; in the majority the position is apical and above the midgut, but in the species of Procornitermes it is displaced and continues into the mesenteric arch (Fig 52).
The members of Syntermitinae have a well-developed mesenteric tongue that is quite varied in shape and may have filiform parts, rounded forms, and a smaller, secondary tongue (Figs 53-55).However, in general the midgut does not possess diagnostic characteristics, except for the species of Silvestritermes belonging to the "lanei group" (S. duende, S. gnomus, S. lanei, and S. minutus), which have a nodule of mesenteric tissue that surrounds the base of the Malpighian tubules [the Termitinae genera of the Termes-Cavitermes group (see below) also have a nodule of mesenteric tissue in a similar position, but it is separated from the Malpighian tubules].number of cushions, and types of ornamentation, and is quite useful to confirm identifications (see Table 1 for references).
The P3 is relatively unimportant as a diagnostic character, although in some species it shows marked features, such as the EVS just after the P2 (tubular or globose), or a partial diverticulum (Armitermes armiger and A. bidentatus).The distal portion of the P3 is generally strongly protracted dorsally, forming a characteristic "U" in most species.The connection between the P3 and P4 can show a distinct isthmus, depending on the genus.The isthmus is inserted apically on the P4 in most of the Syntermitinae, but subapically (Fig 59,arrow) in a very peculiar group of taxa (Acangaobitermes, Noirotitermes, Ibitermes tellustris, Embiratermes parvirostris, E. ignotus, E. brevinasus, and E. silvestrii), which is composed of the smallest species of the subfamily; otherwise the P4 and P5 are not known to be informative.

Termitinae
Termitinae is presently composed of 107 species in 20 genera in the Neotropical Region, and 662 species in 65 genera worldwide (Constantino, 2018).
Except for Cyrilliotermes species and Procornitermes araujoi, all other Syntermitinae have the four Malpighian tubules, with the insertion in two distinct pairs or with the four bases very close together; some species have well-developed ampules (Procornitermes araujoi with four tubules and P. triacifer with two ampules, each with two tubules), but as with the tongue, these structures by themselves have no diagnostic value.
The P1 is dilated, ranging from fusiform to globose depending on the species, and the enteric valve (P2) is always located on the left side of the body.These characteristics are not exclusive to the species of this subfamily: some genera of Termitinae also have the first proctodeal segment inflated, although these can be differentiated from the Syntermitinae by certain combinations of characteristics.The internal ornamentation patterns of the P1 are generally quite characteristic for each species but dissection of this region of the gut is not easy, making them impractical as a diagnostic feature for identification (Rocha & Constantini, 2015).
Although the P2 is always on the left side of the body, the direction in relation to the antero-posterior axis varies greatly among the genera.The connection may be simple and wide (Fig 56) or by means of a constricted portion, forming a loop (Fig 57).The position along the abdomen is a diagnostic feature of some genera: in Cyrilliotermes and Embiratermes sensu stricto (species that form a homogeneous group (Rocha et al., 2012)  Termitinae has constantly been recovered as a nonmonophyletic group (Donovan et al., 2000;Inward et al., 2007;Bourguignon et al., 2014;Bourguignon et al., 2017).This reflects the morphological diversity found in the soldier and worker castes of the taxa that are presently included in this subfamily.In the near future, it is probable that new subfamilies will be designated or re-erected to accommodate some of the genera that are recognized today as members of Termitinae.In view of this likelihood, and the absence of diagnostic characteristics for the subfamily, here we will treat the characteristics that group some of the genera.
Based on the characteristics of the soldiers, the Neotropical Termitinae can be divided into the following groups of genera: the Amitermes group, composed of Amitermes, Gnathamitermes, and Hoplotermes; the Cavitermes-Termes group, which includes the Cavitermes subgroup: Cavitermes, Cornicapritermes, Dihoplotermes, Divinotermes, Palmitermes, and Spinitermes, and the Termes subgroup: Inquilinitermes and Termes; the Neocapritermes group, composed of Neocapritermes, Planicapritermes, and Crepititermes; and the Orthognathotermes group, which includes Dentispicotermes and Orthognathotermes.Some genera, such as Cylindrotermes, Genuotermes, Microcerotermes, and Onkotermes, are so distinct that they cannot be assigned to a group.
In the Amitermes group, the mandibles of the workers are all characteristic of xylophages, i.e., with the apical tooth equal to or smaller than the marginal teeth, and the molar regions with a distinct ridge.In the gut, the stomodeal valve is situated far apically; the mixed segment is very distinct and the mesenteric tongue is long; the P1 is dilated, being fusiform in Gnathamitermes and Amitermes and broader in Hoplotermes; and the P2 inserts on the right side of the body (Figs 60,61).In Gnathamitermes the region of the EVS is differentiated, appearing as a small sac in ventral view.Examples of enteric valves are given in Figs 62 and 63.
The workers of the Cavitermes-Termes group have humivorous mandibles, i.e., with a large apical tooth, much larger than the marginal teeth, and with the molar plate concave and without ridges.The principal characteristic of the gut is the displaced and subapical insertion of the stomodeal valve, and the mesenteric nodule internal to the arch, at the junction with the P1, i.e., at the base of the mixed segment.The enteric valve is inserted on the left side of the body.In the Cavitermes subgroup, the P1 is always tubular, whereas it is more voluminous in the Termes subgroup.The P3 has a fold, which is visible in ventral view, and occurs in all genera except in Cornicapritermes, where the enteric valve is inserted ventrally and the P1 is longer, in the shape of an inverted "C".In the Termes subgroup, the proximal region of the mesenteric tongue is relatively narrow (see Table 1).Examples of enteric valves are given in Figs 64-66.The workers of the Neocapritermes group are heterogeneous in both the mandibles and the gut, but the group was recovered as monophyletic by Bourguignon et al. (2017).A possible synapomorphy is the abscission line on the first abdominal tergite (Fig 67,arrow), present in all species of this group (autothysis occurs in other Termitinae, but apparently the abdomen breaks in other regions).In workers of Neocapritermes, the most obvious character is the pair of protuberances in the distal region of the midgut, which are visible on the right side, and may differ in shape among the species (Fig 68 , arrow).An external feature that can be used to identify several species of Neocapritermes is a row of bristles on the inner part of the foretibia, as is also found in Cornitermes (Syntermitinae).The enteric valve of Neocapritermes opacus is shown in Fig 69.
In Crepititermes the P1 is very short.The workers of this genus may be confused with some small nasutes such as Subulitermes and Coatitermes, although generally the midgut has a characteristic appearance, with more-distinct punctations, which differentiates them from the small nasutes whose midgut has a more homogeneous texture.In workers of Crepititermes the distal region of the P3 has a diverticulum (Fig 70 ), which is a good diagnostic character, since this is the only genus in the Neotropical Termitinae with this feature.The enteric valve of Crepititermes is also diagnostic (Rocha & Cuezzo, 2015).The workers of Planicapritermes have a somewhat flat body, due to their habitat beneath the bark of fallen tree trunks.Refer to the description by Constantino (1998).
The Orthognathotermes group unites species whose workers have mandibles with a marked geophagous pattern, the stomodeal valve is inserted apically, and the mixed segment is very distinct.The P1 is characteristic, long and tubular, passing parallel to the mesenteric arch, with the enteric valve inserted on the right side of the body in dorsal position (Fig 71).The P2 in both genera is very characteristic, with digitiform projections directed toward the lumen of the P3 (Figs 72,73).The workers of this group have a characteristic profile, with the abdomen very globose.
The genus Cylindrotermes is easily recognized by the tubular P1, which forms a clockwise loop (in ventral view) that crosses the abdomen distally, ending on the left side in ventral view (i.e., on the individual's right) (Figs 74,75).The Neotropical Apicotermitinae have a similar loop, but its members are easily recognized by the mandibles, as mentioned above; and when a mesenteric tongue is present, it is situated in the inner part of the mesenteric arch, whereas in Cylindrotermes it is in the outer part.The enteric valve of these species is very difficult to dissect (Fig 76).
The workers of Microcerotermes have xylophagous mandibles and construct an arboreal nest, with covered tunnels passing down the trunk to the ground, very similar to several species of Nasutitermes.Because of this it is common for only workers to be collected, without the soldiers.Although this is not a diagnostic characteristic, the head of some of the workers has a peculiar coloration, ranging from light yellow to brown.The gut is not very informative; the P1 is slightly dilated, with a distal loop; and the P2 insertion is on the left side of the body.None of these characteristics is striking compared to the other Neotropical genera, and the lack of a revision for the Neotropical Region makes their identification difficult.Specimens of Microcerotermes should be carefully compared with the illustrations provided by Sands (1998).The gut of Genuotermes was described in detail by Rocha (2013) and shown to be more closely related to genera of Syntermitinae than to other members of Termitinae (Rocha et al., 2017).The gut morphology is very similar to Cyrilliotermes and Silvestritermes, but with the peculiarity that the insertion of the isthmus in the P4 is subapical.
Onkotermes was described by Constantino et al. (2002), who transferred the species Amitermes brevicorniger to the new genus, basing this, among other arguments, on the morphology of the gut, which is highly peculiar in this genus.Members of Onkotermes are found in open areas in the southern Neotropics (Pampas and Chaco Seco regions), and it is highly uncommon to find soldiers in collections, which makes identification of the workers particularly important.In the workers, the mandibles are highly developed for xylophagy, the mesenteric arch shows the mixed segment convexly arched, and the P1 is inserted in a well-marked EVS in the P3 (Torales & Fontes, 2008).

Final considerations
The ability to identify termites using only the worker caste is increasingly necessary.For example, various ecological studies use collection methods (using soil monoliths or termite mounds) that indiscriminately sample the diversity of termite species present, often represented only by the workers.This contribution summarizes the recent advances in the taxonomy of Neotropical termite workers.Modern descriptions of taxa treat the worker caste in detail, although many questions remain to be answered.Little is known about the internal anatomy of pseudergates, especially in the Kalotermitidae, and it is possible that other taxonomically useful features exist, that have not yet been explored, including a better study of the mandibles or gizzard (Myles, 1997).Also, there is a need to continue and expand comparative studies of members of Termitidae, particularly for Termitinae and Apicotermitinae.As noted above, the Termitinae is not a natural group, and investigating the internal anatomy of workers based on established characters, or possibly new characters, will help in analyses of the relationships and evolutionary history of the species, and will also help in the rapid identification of pest or ecologically important species.Further studies of the internal morphology of the various parts of the gut may help to test the hypotheses on fixation and movement of intestinal symbionts.Outside of the field of taxonomy and systematics, it is necessary to encourage research on the actual diets, such as analyses of gut contents, of several species of Termitidae, to enable comparison with the fine morphology of the mandibles and the gut.
For all the subfamilies of Termitidae the number of specific studies on feeding behavior is insufficient, and current hypotheses are based more on the morphology of the mandibles and characteristics of the gut anatomy of workers, as well as information on the collection site.To better understand the true food sources of many species of Termitidae, more effort is needed, including studies of gut contents, which can provide clues to the food type.
The present contribution is an attempt to summarize a broad array of characters, so that the process of screening and identifying termites using the worker caste can be decentralized from the hands of experienced taxonomists and made useable by any interested researcher.In the near future, it is hoped that every termitologist, even with little experience, will be able to identify termites regardless of the availability of the soldier or imago caste.
defined the absence(Fig 22)   or presence of the mixed segment(Fig 23); or even vestigial mixed segments, in which there is no proper mesenteric tongue, but an oblique portion of the midgut.The joining of the P1 (through the P2) with the P3 may be visible in dorsal view, right lateral view (and in this case can be seen both dorsally and ventrally), or in ventral view.The shapes of EVS often vary.Both the gizzard and the insertion of the Malpighian tubules in the Neotropical Apicotermitinae have been little studied.

Fig 27 .
Fig 27.Worker leg of Apicotermitinae sp., one of two rows of tick bristles on tibia.
fig. 4A, B; and a new species currently being described by JPC, Fig 28).The same can be said of some species within the same genus, such as Ruptitermes (figs 24F and K, 25A and E of Acioli & Constantino, 2015) and Aparatermes (Figs 29 and 30); or even between species of different genera, such as Ruptitermes franciscoi (fig.24 J-K of Acioli & Constantino, 2015) and Aparatermes cingulatus (Fig 30), or Ruptitermes bandeirai (Fig 31) and Tetimatermes oliveirae (Fig 32), where the scale distribution and/or spine ornamentation is very similar.Also, it is crucial to consider the existence of intraspecific variations, such as those observed in Anoplotermes janus (compare fig. 2E of Bourguignon et al., 2010 and Fig 33) and Anoplotermes pacificus (Figs 34, 35, JPC and TFC, personal observation).

Fig 50 .
Fig 50.Foretibia of Cornitermes cumulans worker.Certain species or groups of species of Syntermitinae show marked synapomorphic characteristics of the internal anatomy.The gizzard and crop generally follow the basic plan for the Termitidae (as described by Noirot, 2001), but Ibitermes curupira, Embiratermes robustus, and the species of Mapinguaritermes show a very characteristic ornamentation of spines at the base of the first-order pulvilli (Fig 51).

Fig 68 .
Fig 68.Neocapritermes opacus worker gut in situ, right view (arrow: bilobate protuberance in the distal region of the mesenteric tongue).

Fig 76 .
Fig 76.Enteric valve armature of Cylindrotermes sapiranga partially dissected, the musculature is partially removed, one of the three cushionsis highlighted in the figure.

Table 1 .
List of taxonomic studies that treat the worker caste of Neotropical Termitidae genera.(Continuation)