798 On the Mound of Macrotermes michaelseni as an Organ of Respiratory Gas Exchange J. Scott Turner* Department of Environmental and Forest Biology, College of Environmental Science and Forestry, State University of New York, Syracuse, New York 13210 Accepted 7/6/01 ABSTRACT Patterns and rates of air movements in the mounds and nests of Macrotermes michaelseni were studied using tracer methods. Wind is a significant source of energy for powering nest ven- tilation, despite the mound being a completely enclosed struc- ture. Nests are ventilated by a tidal movement of air driven by temporal variation in wind speed and wind direction. Density gradients sufficiently steep to drive bulk flow by natural con- vection will be rare. However, metabolism-induced buoyant forces may interact with wind energy in a way that promotes homeostasis of the mound atmosphere. Introduction Mounds built by termites of the family Macrotermitinae are a prominent feature of the tropical savannas of southern Africa (Harris 1956; Kalshoven 1956; Lu ¨scher 1961; Ruelle 1964; Ruelle et al. 1975; Pomeroy 1977; Collins 1979; Darlington 1984, 1985). The colony that constructs the mound comprises as many as 2 million individual termites. Surprisingly, the mound is not a habitation but is simply the most visible com- ponent of a structure that extends well below the ground. The mound is not a haphazard pile of spoil from excavation of the nest either. Within the mound and surrounding the nest is an extensive and stereotyped network of air spaces (Fig. 1; Dar- lington 1985; Turner 2000a). The complex architecture of the mound implies some phys- iological function, and prevailing opinion has long been that the mound functions to regulate the nest environment (Lu ¨scher 1956, 1961; Wilson 1971; Darlington et al. 1997). In the late 1950s, the Swiss entomologist Martin Lu ¨scher proposed that * E-mail: jsturner@mailbox.syr.edu. Physiological and Biochemical Zoology 74(6):798–822. 2001.  2001 by The University of Chicago. All rights reserved. 1522-2152/2001/7406-00140$03.00 the mound of Macrotermes bellicosus (misidentified by Luscher as Macrotermes natalensis; Ruelle 1970) functions essentially as a colonial heart-lung machine (Lu ¨scher 1956, 1961). In this conception, the colony’s high metabolic rate, which by some estimates runs into the hundreds of watts (Darlington et al. 1997), heats and humidifies the nest air, reducing its density. The resulting buoyant forces circulate air through the nest and the surface tunnels, which are sites for exchange of heat and respiratory gases. Homeostasis of the nest environment follows from a linkage between circulation rate and metabolism. Higher rates of metabolism supposedly impart greater buoyant forces to the nest air, which would in turn drive a more vigorous circulation. This mechanism has been called “thermosiphon ventilation,” although a more accurate designation would be “metabolism-induced natural convection.” Lu ¨scher’s (1956, 1961) ingenious idea enjoys widespread ac- ceptance today but not because there is positive evidence sup- porting it. The postulated thermosiphon flows were inferred from distributions of temperature, humidity, and oxygen con- centration measured within the mounds, and metabolism- induced natural convection is but one way to explain his results. Consequently, Lu ¨ scher’s (1956, 1961) thermosiphon model has, since its inception, been criticized for failing to account for the complex variation of mound architecture among the macro- termitines and for the interactions of the mound with wind and other aspects of the physical environment, notably tem- perature (Loos 1964; Ruelle 1964; Korb and Linsenmaier 2000). The thermosiphon model makes testable predictions, how- ever, which can be falsified if large-scale patterns and rates of airflow within the mound and nest can be measured. This article reports such measurements in the nests and mounds of Macrotermes michaelseni, a widely distributed species that builds enclosed mounds similar to those of M. natalensis and M. bel- licosus. I have found that, while aspects of the thermosiphon model have merit, patterns and rates of airflow and gas exchange in these mounds are far more complex than those predicted by Lu ¨scher. These data clarify how the mounds func- tion as organs of external physiology (sensu Turner 2000b) and how social homeostasis emerges from the complex architecture of the mound. Material and Methods Species The subject of this study is the southern African Macrotermes michaelseni Sjo ¨stedt, formerly Macrotermes mossambicus Holm-