Behavior of the North American termite Tenuirostritermes tenuirostris, with special reference to the soldier frontal gland secretion, its chemical composition, and use in defense

Tenuirostritermes is a small genus of Neotropical termites with two species ranging northward into warm temperate, semi-arid areas of southwestern Texas [cinereus (Buckley)] or beyond into southeastern Arizona [tenuirostris (Desneux)]. As a member of the Nasutitermitinae, it is somewhat more primitive than the very large and well-known tropicopolitan genus, Nasutitermes. The subfamily, containing the most highly specialized members of the Termitidae, is the largest in the order Isoptera. The major specialization within the subfamily has involved gradual reduction of the soldier mandibles to non-functional stubs with concomitant development of a small projection on the front of the head into a long, attenuated snout or nasus. The frontal gland, occupying most of the bulbous posterior of the head capsule, elaborates a defensive secretion which can be forcibly ejected through the frontal pore at the tip of the nasus by contraction of powerful mandibular muscles. This fontanellar gun represents the apex of sophistication among the varied chemical defense mechanisms of the termites. The zoogeography and affinities of the nasutitermitine genera have been discussed by Emerson (1955) and Krishna (I97o) and their defensive behavior by Ernst (I959), Moore (I969) and Wilson (I97 I). Light and Weesner (I955)


INTRODUCTION
Tenuirostritermes is a small genus of Neotropical termites with two species ranging northward into warm temperate, semi-arid areas of southwestern Texas [cinereus (Buckley)] or beyond into southeastern Arizona [tenuirostris (Desneux)]. As a member of the Nasutitermitinae, it is somewhat more primitive than the very large and well-known tropicopolitan genus, Nasutitermes. The subfamily, containing the most highly specialized members of the Termitidae, is the largest in the order Isoptera. The major specialization within the subfamily has involved gradual reduction of the soldier mandibles to non-functional stubs with concomitant development of a small projection on the front of the head into a long, attenuated snout or nasus. The frontal gland, occupying most of the bulbous posterior of the head capsule, elaborates a defensive secretion which can be forcibly ejected through the frontal pore at the tip of the nasus by contraction of powerful mandibular muscles. This fontanellar gun represents the apex of sophistication among the varied chemical defense mechanisms of the termites. The zoogeography and affinities of the nasutitermitine genera have been discussed by Emerson (1955) and Krishna (I97o) and their defensive behavior by Ernst (I959), Moore (I969)  shrub-invaded desert grassland ecotone, characterized by scattered trees, shrubs and cacti. Four groups were taken from under stones during one afternoon and three nearly complete foraging groups were taken on the soil surface near midnight. Termites and a minimum of soil were quickly scooped up with a trowel and placed in a large enameled tray. A few hours later the termites were separated from soil and debris by sifting and hand-picking. To determine soldier/ worker ratios, these castes were counted in each of the two combined collections, those from under stones, and those taken on the surface. Several hundred of these soldiers and wo.rkers were reserved for chemical analyses.
A small foraging group was collected later for behavioral studies.
They were supplied with fine, dry native vegetation and survived reasonably well in the laboratory for about three weeks in a plastic petri dish of moist soil. A few minor workers of the myrmicine ant, Pheidole desertorum Wheeler, were also collected in the same area and maintained for the observation of termite-ant encounters. Although most species of Pheidole are reportedly seed gatherers rather than habitual predators, this was apparently of no consequence in the artificial situations described below.
No encounters between termite foraging groups and small arthropod predators were seen in the field. A few random encounters were staged in the laboratory by introducing individual ants into the dish containing the small group of termites. The defensive behavior of the soldiers was observed in detail under low magnifications of a stereomicroscope, during individual soldier-ant encounters contrived as follows: Single living ants were mounted in a natural position on glass slides with a small drop of rubber cement. Slide and ant were then placed on the soil in the observation chamber with the termites.
In this distinctly one-sided situation, one or more soldiers invariably located and attacked the ant within a few minutes.
Attacks by five different soldiers were quite literally recorded on 1974] Nutting, Blum, & Fales Tenuirostritermes 69 separate slides in the form of discrete threads of frontal gland secretion which they had fired in the general direction of the ants. The volume of most of these shots was calculated from measurements made of the threads with an ocular micrometer under a compound microscope. Drawings to illustrate the variation of individual shot patterns were made with the aid of a camera lucida on the same microscope. The fate of a few ant victims was followed until their death and compared with that of normal ants which were isolated and allowed to die of starvation and desiccation.
For the chemical analyses 60o living soldiers were placed in reagent n-pentane. Although it had been determined that the' heads of the soldiers were the primary source of the odorous defensive compounds, 400 living workers were also extracted for comparative purposes. The extracts were concentrated under vacuum and analyzed on a gas chromatograph interfaced to an LKB-9ooo mass spectrometer. A 4-meter column of o% SP-Iooo on 8o/oo mesh Supelcoport operated isothermally at 65C was employed for all analyses.

RESULTS
Foraging behavior.
T. tenuirostris forages in the open on the soil surface, mainly at night in southern Arizona as we have recently established, but it is also active on cloudy days, at least in western Mexico (Nutting, I97O). The following account of its general foraging behavior is a synthesis of observations made on about 2 different groups at the Santa Rita Experimental Range during September-October, 1972. Workers presumably open one or more access holes to the surface, but the soldiers are generally the first out and the last to. return underground during bouts of surface activity. As many as OO or more soldiers and very few workers may congregate within a 5-to Io-cm radius of the hole prior to. foraging. Little organization is at first apparent except that many of the soldiers may be standing on the alert with heads pointed peripherally. One or more discrete foraging columns eventually emerge from such unorganized, milling masses, perhaps after the soldiers are joined by a critical number of workers. Once in motion, columns up to. six workers wide move quickly in an ant-like manner, and show very tight trail-following behavior. While soldiers move. out with the workers, they eventually take up stations at I-to 2-cm intervals on either side o.f, and ca. cm Irregular craters of soil, up to 8 cm in diameter and cm high, are often thrown up around the holes during periods of surface activity. Ground-level runways radiating from the hole may be bridged over to transform a crater into a rough pile of soil 2 to 3 cm high. Since individu.als of this species are relatively small (length of soldier, 3.o ram; of worker, 4.5 mm), access holes as large as 5 mm in diameter probably provide for efficient, emergency re-entry of large numbers of foragers in a very short time. Workers have been observed to close the exposed holes shortly after the conclusion of surface activity and, although a few soldiers remain in and around the hole during closure, none was ever left outside. The same holes are apparently used repeatedly, at least for periods of several weeks.
The high proportion of soldiers in the foraging groups indicates the relative importance of this caste in those species of termites which routinely forage on exposed trails. The four combined groups collected from colonies beneath stones contained 545 individuals with a soldier:worker ratio of I:3.82 (20.7% soldiers). Three combined foraging groups contained 1,549 individuals with a soldier:worker ratio of :I.23 (44-9% soldiers). A single oraging group of 230, perhaps recently .organized, had a ratio of soldier to o.4o worker 71.3 % soldiers).
Soldier behavior. Whenever the laboratory group of soldiers and workers was disturbed by jarring or removing the cover of their container, several soldiers, with antennae raised and waving, congregated on the edges of objects and on elevated vantage points. Individual workers of the ant, Pheidole desertorum (ca. 3.5 mm long), introduced into this situation were quickly fired on at very close range by several soldier termites in rapid succession. The appendages of such victims were almost completely immobilized, within o sec or less, by the viscous, glue-like threads thrown over them.
Further details of soldier defensive behavior are based on observations made during encounters with living ant workers fixed to glass slides. Although blind, the soldiers presumably orient towa.rd such threats by olfactory or auditory cues. An alerted soldier closes rapidly on any potential target and may actually touch it for an instant with its antennae before firing from a distance of ca. o.5-3.o ram). As it fires, it may jerk forward or backward, in position, and may or may not then oscillate to-and-fro one or more times. This apparently insignificant bit of behavior greatly enhances the effectiveness of the soldier's shot by throwing a bight, or even one or more loops, in the thread ( Fig. ). Attacks by as many as six soldiers from different directions were invariably fatal to an ant. Soldiers often examined victims with their antennae for up to a minute, occasionally becoming temporarily entangled themselves. They were never seen to wipe the nasus on the substrate after any of this activity.
In unstaged encounters, soldiers usually fired but once at an ant; however, by carefully prodding and exciting the fixed ants, individual soldiers were induced to fire from one to six additional shots at them. Data on the shots recorded on and near four such ants are summarized in Table I. It was impossible to determine the order of shots fired by individual soldiers, but those of greater diameter and length are probably the ones fired at or near the beginning of an encounter.
Longer shots often began with a series of droplets, and occasionally contained short breaks in the thread. It would have been interesting to determine what percent of their glandular holding capacity had been spent in these individuals, but drought conditions during seasons of activity have since prevented further collections.
Although the ant victims were physically immobilized within a few seconds, they also appeared to be adversely affected by some of the chemical components of the soldier secretion, probably the terpenes. 1974] Nutting, Blum, Fales Tenuirostritermes 173 They lay very quiet, moving their appendages but little even when prodded. Their struggles were very feeble after 1.5 hr, while only occasional, slow contractions could be elicited after 4 hr. They appeared essentially dead after 5-6 hr. By comparison, normal ants, similarly confined without food or water, remained so for 5-1o hr, showed postural difficulties after O-I2 hr, responded vigorously to prodding for as long as 24 hr, and did not appear to be dead until 24-33 hr after confinement.
Chemical Analyses. Three compounds accounted for over 90% of the low boiling volatile compounds which were detected in the extracts of the soldiers. Based on the congruencies of their gaschromatographic retention times and mass spectra, the volatiles were identified as a-pinene, myrcene, and limonene. The mole percentages of the compounds were: a-pinene 62%, myrcene 27c, and limonene I%. None of these monoterpene hydrocarbons was detected in the extracts of the workers. Ernst (1959) was inclined to believe that the frontal gland secretion of soldiers of a Nasutitermes sp. from the Ivory Coast was essentially non-toxic and that it functioned primarily to entangle its enemies. He did not rule out the possibility of .an insecticidal effect.

DISCUSSION
Considering that normal worker ants survived without food or water Similarly, with N. costal# they ound a-and fl-pinene and limonene among six major constituents o the secretion, and myrcene again among three minor components. However, although it is evident that monoterpene hydrocarbons are rather characteristic defensive products o nasute soldiers, these compounds do not appear to ha,ve a widespread distribution in the Insecta. Indeed, with the exception o the myrmicine ant Myrmicaria natalensis F. Smith, which produces limonene in its poison gland (Grfinanger et al., 96o), monoterpene hydrocarbons have not been identied in any other insect taxa.
By taking high speed motion pictures o Nasutitermes soldiers attacking vestigial-winged Drosophila, Ernst (959) determined that they move rapidly orward and backward (only once ?) in delivering the jet o, rontal gland secretion. He also observed one soldier to move its head rom side to side during an attack. Our tracings o the shot patterns ( Fig. ) show that either or both o these movements are also used by soldiers o T. tenuirostris to increase the effectiveness o their defense. The basic oscillatory behavior pro.bably represents the jerking or rocking movements which termites commonly exhibit in mild alarm situations. We have also shown that soldiers are capable o/]ring more than once, i sufficiently alarmed, and certainly do not deliver the entire contents o the fro.ntal gland in a single discharge as Ernst suggested.
Many authors have stated that ants are the chie predators o social insects, with the termites not the least among their prey (Wilson, I97)-There can be little doubt that the nasute soldier is a highly effective deender o the colony against arthropod predators close to the size range o the termites themselves. A large orce o nasutes would seem to be o critical importance in protecting the workers on their food-gathering expeditions, particularly those that forage in the open as does T. tenuirostris. However, this common assumption apparently rests largely on casual, subjective observations and circumstantial evidence o the type presented here. Quantitative data required to prove this assumption are not available and would be extremely difficult to obtain. 1974-1 Nutting, Blum, & Fales--Tenuirostritermes 175 The high proportion of soldiers maintained by T. tenuirostris, and presumably many other free-foraging termites, certainly attests to their great importance in the colony. Our figure of 2o% soldiers in combined collections from four different colonies is in fair agreement with Light and Weesner (1955) who determined that incipient colonies of this species usually contained between 25 and 33% soldiers. Three of our foraging groups combined contained 44.9% soldiers, while a single group contained 71.3%. Ernst (I959) reported 7o to 9o% soldiers in outer portions of his laboratory colony of Nasutitermes sp. from the Ivory Coast. Although demands on the defensive capabilities of the soldiers underground must be trivial or rare, indeed, these figures show that a very high proportion of soldiers is mobilized for foraging parties.
This situation is in striking contrast to that in another subterranean termitid, Gnathamitermes perplexus (Banks) (Amitermitinae), which is very common over the range of T. tenuirostris in southern Arizona. This species forages widely on the surface under cover of thin, soil sheeting which it builds over all types of dead plant material (Collins et al., 1973). Based on 74 foraging groups, the average soldier:worker ratio is :8 9 (I.II% soldiers). Individual groups ranged from to IO3I termites and contained from o.o to 9.o% soldiers (Nutting and Haverty, unpublished).
During the few hours of field observations on the above-ground activities of r. tenuirostris, no encounters with ants or other small predators were seen. However, the presence of an impressive ant tauna in our study area lends further credibility to the importance of the soldier caste in the economy of this termite. In a preliminary study of the ants on the Santa Rita Experimental Range, Gaspar and Werner (unpublished) list 33 species, with high populations of Crematogaster coarctata vermiculata Emery and Forelius foetidus (Buckley), and estimate I753 nests/ha for all species. About one fourth of the species are regular predators, while at least an additional third are occasional predators or at least feed on animal matter. Many of them forage at night during the summer rainy season when Tenuirostritermes is also most active on the surface. Although we have no estimates of colony size or density for the termite, its colonies are fairly common in the area and must contain several thousand individuals.