Interpretive notes on the termite family Hodotermitidae in Africa, part 1: Hodotermes (Coaton and Sheasby, 1972, 1975)

@tonyrebelo @beetledude @botswanabugs @jtermiteo @r_r_r_ @hamishrobertson

The aim of this Post is to provide an interpretive review of several half-century old, but still best-of-their-kind, papers about the African members of an ecologically extremely important family of insects, viz. Isoptera: Hodotermitidae.


My first reference is Coaton W G H and Sheasby J L (1975,

Hodotermes is restricted to subSaharan Africa, and contains two spp.

Hodotermes occurs in southern and East Africa, as far north as Eritrea. However, H. mossambicus occurs only as far north as Ethiopia. Hodotermes erythreensis replaces H. mossambicus in Eritrea, Somaliland, and possibly Somalia and easternmost Ethiopia (

My commentary:
Biogeographically, H. mossambicus is remarkable in paralleling a pattern in the ungulates, in which

  • many genera and spp. occur in southern and East Africa, but not West Africa,
  • there is a broad disjunction between southern and East Africa, and
  • the Horn of Africa tends to have a distinctive fauna.

Hodotermes mossambicus occurs mainly where mean annual rainfall is <750 mm, except for East Africa, where it may extend to 800 mm. It occurs widely in Namibia (50-600 mm) and Botswana (200->600 mm), including the southwestern Kalahari (Kgalagadi Transfrontier Park,

Hodotermes mossambicus is absent from most of the Namib, but present in the Namib in Kaokoland and at Walvis Bay. In the Kalahari, it is restricted to the finer-grained substrates, and perhaps absent from the coarsest, nutrient-poorest sands. However, it is present (see Coaton 1963, and in sparse populations on dunes/sandflats.

My commentary:

Hodotermes mossambicus is somewhat tolerant of salinity (also see Popov et al.). In Zimbabwe, it is absent from granitic substrates, except where saline.

My commentary:
Hodotermes mossambicus seems to prefer base-saturated substrates.

Each colony has several/many hives (diameter <60 cm), at densities of 85-96 hives per hectare. The hives are 1.8-12.2 m deep. Only a few of the hives in each colony contain reproductives. Storage chambers for food are extraneous to the hives, and are mainly <30 cm deep. The hives within each colony are linked by tunnels, 2.5-7.5 cm in diameter.

My commentary:
The subterranean location of the hives, with no clue at the surface, protects H. mossambicus from predation by large myrmecophages ( This helps to balance the exposure of the workers as they scurry on the surface (including in cold weather) while foraging.

The diet is mainly, but not exclusively, grass.

My commentary:
In contrast to the normal conception of a termite, Hodotermes mossambicus acts as a grazer of mainly dry 'sweet' grass. It drinks at depth, as opposed to the horizontal commuting performed by grazing ungulates.


In Zululand, the mean annual production of grass is at least 3 tonnes (dry matter) per hectare. Hodotermes mossambicus is capable, in its own right, of harvesting most of this production.

My commentary:
This implies suppression of wildfire, and conservation of nutrient elements, e.g. nitrogen, sulphur, and selenium, that are easily volatilised and depleted by combustion.

Coaton and Sheasby (1975) state:
"Coaton (1958) reported that extensive areas in Zululand carried saturated populations of H. mossambicus which were consuming grass at an estimated rate of 1-3 tons per morgen (1.06-3.17 metric tons per hectare) per year." Some populations consumed the entire crop of hay on a given site, where production of hay is about 3 tons per morgen (3.17 metric tons per hectare).

"Veld hay was produced at a rate of just on 3 metric tons/hectare over stretches of Lowveld and Zululand Thornveld in the Game Corridor between the Hluhluwe and the Umfolosi Game Reserves, an area not grazed by livestock in which H. mossambicus is present, although not in pest extreme case is cited of an ungrazed area in Zululand, which for years had been completely denuded by the end of winter, yielding 3 metric tons of baled hay per hectare per annum after saturated harvester termite infestation had been exterminated; at the other end of the scale, a far lighter infestation in an ungrazed test enclosure in the central Orange Free State has been recorded as removing only about 100 kg per hectare per year out of a total hay production of just over one metric ton".

My commentary:
I infer that H. mossambicus is fully a part of the habitat of wild ungulates in Zululand, even though this area is on the moist side, climatically. The map in Coaton and Sheasby (1975) shows the species to be present in Ndumo. uMkhuze, and St Lucia reserves. This termite shares a guild with the fauna of ungulate grazers of Zululand, which are noteworthy in that they extend from tropical to subtropical latitudes.

"On numerous occasions foraging workers of H. mossambicus have been observed to be cutting down lush green grass, the supplies gleaned being taken directly underground or left lying on the surface around the foraging ports, for subsequent collection...Thriving colonies have also been recorded as doing extensive damage to growing small-grain, legume and fodder crops where the diet must have consisted almost exclusively of green plant material."

Coaton and Sheasby (1975) describe a situation in which the entire diet of a colony for years was lawn grass, viz. Cenchrus clandestinus ( in a green state.

"From the observations presented above it is clear that under natural conditions in the field colonies of H. mossambicus can thrive on a practically undiluted diet of green vegetation" (appropriately cured, of course, after cutting).

My commentary:
This is a clear reference to

  • foraging by day in winter, and by night in summer,
  • damage to agricultural crops, and
  • consumption of not only dry but also green grass, the latter being in some cases 100% of the diet, and/or 100% of the green grass present on a given site.

My second main reference is Coaton and Sheasby (1972,

Referring to Namibia, these authors state:
"Including the vast Northern Kalahari region, which is rather lightly infested by harvester termites except on the more consolidated soils of river flood-plains, omurambas, vleis, and dune valleys, it is very conservatively estimated that H. mossambicus consumes approximately 25% of the aggregate hay yield of South West Africa during every year of average rainfall. In years of drought, when grass regeneration is poor, this figure would be doubled or even trebled. Apart from lowering the stock-carrying capacity of the territory by at least one-quarter these termites are facilitating soil erosion and desert encroachment through their removal of the protective grass cover."

The following seems to have no presence on the Web, even as a mere citation:
Grube S (2000) Soil clumps - indicators of foraging activity by Hodotermes mossambicus (Hagen) (Isoptera, Hodotermitidae) in northern Namibia. Cimbebasia 16: 269-270.

This author states that dumping of soil at the surface, by Hodotermes mossambicus, is not correlated with foraging activity. Foraging activity peaked in the rainy season, whereas the dumping of soil peaked in winter.

My commentary:
Are these termites feeding fertiliser to the grass and herbivores? Why is it that no author mentions the faeces of Hodotermitidae, which - in my experience - appear dark and nutrient-rich, like valuable fertilizer?

Lee and Wood (1971,, on page 180, state in a worldwide context: "It is only in South Africa that the depredations of termites in pastures have been considered sufficiently injurious to merit the adoption of extensive control measures."

However, even in central Asia, the hodotermitid Acanthotermes ahngerianus is a pasture pest, where grass grows. Bare areas around the nests may occupy 20% of the surface. (Also see the second comment below.)

My commentary:
I take this as evidence of the role of herbivory - with ungulates and termites working in concert - in producing low, palatable, fire-free vegetation in South Africa. I note that no termite on Earth is known to cut and maintain hedges or lawns. The fact that H. mossambicus eradicates lawns (patchily), rather than promoting or maintaining them, is crucial for understanding how the same species can be despised as a pasture pest by pastoralists, and valued as a natural member of the grazing guild by wildlife managers who understand the longer-term, more holistic relationships among herbivory, fire, and nutrient-cycling.

to be continued in

Posted on 05 June, 2023 02:28 by milewski milewski


Do Hodotermes or Microhodotermes have similar effects to the following?

Kozlova A V (1951) Accumulation of nitrates in termite mounds in Turkmenia. Pochvovedenie 10: 626-631

Lee and Wood (1971) state on pages 109 and 174:

"Accumulation of nitrates in mounds of Anacanthotermes ahngerianus ( in Central Asia was noted by Kozlova (1951). Mounds contained 0.85% NO3 compared with 0.022% NO3 in surrounding soils. Kozlova estimated that the mounds (>500/ha) contained 208 kg/ha of NO3 or 47 kg/ha of readily available nitrogen, and considered that the favourable conditions of temperature and moisture in mounds favoured mineralisation of nitrogen contained in the organic matter in mounds. The mounds are used by farmers as fertilisers in the initial stages of agricultural development.

"This activity by termites may help to explain the great concentrations of nitrogen in the nitrate solonchaks and takyr soils ( in these parts of central Asia.

Please note that most of Turkmenistan lies at latitudes slightly higher than those reached at the southern tip of Africa. This means that the solar radiation received by the hodotermitid in this area is less than that received in southernmost South Africa, by Microhodotermes, and winter is colder.

Posted by milewski 6 months ago

The aardwolf (Proteles cristatus, is the largest myrmecophage on Earth that is specialised for a diet of termites (

Its diet is mainly Trinervitermes and Hodotermitidae, ostensibly in that order of importance.

However, its distribution ( corresponds better to that of Hodotermes (both spp.) and Microhodotermes viator than to Trinervitermes.

Trinervitermes is present in savanna in West and west-central Africa, where the aardwolf is absent. However, does Trinervitermes extend into the Sahel (, where the aardwolf is also absent?

Posted by milewski 6 months ago


Jeremy Midgley ( and told me of an observation during fieldwork on 25 July 2001, in fynbos on the property Elandskloof, on the lower slopes near Sir Lowry's Pass (

The substrate was brownish or reddish soil, not whitish sand. The vegetation consisted of Protea repens (, Ericaceae, Restionaceae, and other taxa typical of fynbos. Dicerothamnus rhinocerotis was absent, so this was not 'renosterveld' or transitional to 'renosterveld'.

There are no heuweltjies ( anywhere near this location.

The weather was a sunny, mild day in winter.

The stands in question had been recently burned, with the regenerating plants still only a few cm high.

Microhodotermes viator was noticeably active on the surface, at several locations. The groups of workers were collecting plant material, consisting of not only litter but also green parts, cut by the insects themselves.

I surmise that mean annual rainfall at this location is about 750 mm (

Jeremy Midgley also told me that he had previously observed M. viator in 'pure fynbos' under mesic conditions at other locations.

My commentary:
The closest observations in iNaturalist are and and However, also relevant are and What I find noteworthy is that M. viator here penetrates ecosystems so nutrient-poor (partly owing to leaching) that they have negligible utility for farming. The indigenous fauna of mammalian herbivores consists mainly of Taurotragus oryx, Pelea capreolus, and Raphicerus melanotis (all in sparse or transitory populations), plus Otomys sp. (at least in mature stands). Microhodotermes viator is among the few animals that can survive on the fibrous, flammable plant matter growing on such sites, and should be included in the same ecological guild as the above mammals.

Posted by milewski 6 months ago

In southern Africa, the aardwolf (Proteles cristatus) occurs as far east as southwestern Mozambique. However, in iNaturalist there are currently no observations of this species in Kruger National Park.

According to Pienaar (1964,

The aardwolf is actually fairly rare in Kruger National Park, being more or less restricted to the vicinity of Pretoriuskop (, where the soils are relatively nutrient-poor. It has only occasionally been recorded on the nutrient-rich basaltic substrates, bearing open vegetation, near e.g. Tshokwane (

By contrast, the aardvark (Orycteropus afer, is widespread and common in Kruger National Park (Pienaar, 1964), and has been observed here several times in iNaturalist.

Posted by milewski 6 months ago


Patrick Hurley told me that, on 29 January 2000, he observed astonishingly large numbers of what I take to have been either Microhodotermes viator or Hodotermes mossambicus, in the farmyard of the farm Kleinplasie (registered name Kleinplaats), between Bredasdorp ( and Klipdale (, in the Western Cape province of South Africa.

The location is 34 degrees 24 minutes S, 19 degrees 59 minutes E.

The farm in question is devoted to Ovis aries and Triticum (, and does not retain much natural vegetation.

This observation indicates the episodicity of activity in hodotermitids. More particularly, it indicates either the ability of M. viator to persist in anthropogenically altered environments, or the anthropogenic spread of H. mossambicus into the winter-rainfall climate.

Because Patrick Hurley described the congregations in the farmyard as 'plague-like', I suspect that the latter is applicable.

Patrick Hurley also reported that Otocyon megalotis megalotis (, which preys on hodotermitid termites, had become so common on and near this farm that it was by then regarded as a pest. It excavates the walls of farm dams, causing them to collapse.

This semi-myrmecophagous species of canid may possibly have increased in recent times. If so, part of the reason for this facilitation might be an anthropogenic extension in the range of H. mossambicus, in response to extensive farming of wheat.

Posted by milewski 6 months ago

Stuart et al. (2003, studied the diet of Otocyon megalotis megalotis.

The study site was at the semi-arid edge of Fynbos Biome, with winter rainfall.

These authors found that a major component of the diet was Hodotermes mossambicus.

However, it remains possible that the real identity of this termite was Microhodotermes viator.

What reduces my confidence in the identification presented is that these authors do not even mention M. viator, despite this being the species expected in this area.

On the other hand, if the identification was valid, then this would suggest anthropogenic facilitation, and westward extension, of H. mossambicus in response to farming.

Posted by milewski 6 months ago

Lovegrove B G (1991) Mima-like mounds (heuweltjies) of South Africa; the topographical, ecological and economic impact of burrowing animals. Symp. zool. Soc. Lond. 63: 183-198.

Heuweltjies have diameter about 30 m, height <2 m. This is a good reference for the fertility of heuweltjies. The soils on the heuweltjies are nutrient-rich, water-holding, organic, alkaline, and clayey.

Posted by milewski 6 months ago

Coaton W G H (1937) The harvester termite: the biology, economic importance and control. Farming in South Africa June 1937, pages 249-252.

Hodotermes mossambicus often attacks houses (wallpaper, books, carpets, curtains). It consumes not only lawns, but also horticultural shrubs and saplings. It destroys annual herbaceous crops.

Hodotermes mossambicus is absent from "bare rocky mountain areas of the Drakensberg and in the very damp eastern coastal regions of Natal, the Transkei and Pondoland".

This species prefers well-drained slopes for its subterranean nests. The hives are situated 0.6-6.1m below the surface.

A peculiarity of H. mossambicus is that the queen remains active. It commutes among the several hives of the colony, and remains able to flee from danger.

My commentary:
I note the following difference from fungus-culturing ants, in the Americas. Hodotermes mossambicus is available to termite-eating animals on the surface without the defences of the soldier caste, because the soldiers remain below ground. Soldiers never appear on the surface, ostensibly because they have a specialised role in protecting the colony from predacious ants underground. Furthermore, the workers exposed on the surface include juveniles, which are particularly palatable. I take this vulnerability as a corollary of the remarkable productivity of H. mossambicus, relative to most termites.

"The fungus-grower termite species such as Macrotermes natalensis and Odontotermes badius are well represented in the Transvaal as well as the Orange Free State, and in the absence of wood, these species are found to live mainly on dry grass. As the grass is mown down by the harvesters, the fungus growers cover the stubble with clay, under cover of which the plants are destroyed down to the roots".

My commentary:
This is evidence of the importance of fungus-culturing termites, in combination with Hodotermes, in the Highveld.

Posted by milewski 6 months ago

Coaton W G H (1981) Fossilised nests of Termitidae (Isoptera) from the Clanwilliam district, Cape Province. J. ent. Soc. sth. Afr. 44(2): 79-81

Anacanthotermes (Hodotermitidae) occurs as 4 spp. in the Oriental realm, and 7 spp. in the Palaearctic realm. This genus does not occur in the Ethiopian realm.

Posted by milewski 6 months ago

van Ark H (1961) Harvester termite control in the Karoo. Farming in S. A. 37(7): 46-48.

"the most serious insect pest in the Karoo"

My commentary:
...even compared to locusts?

H. mossambicus does extensive damage to maize, winter cereals, and lucerne.

When hungry enough, H. mossambicus consumes woody materials, e.g. shelves and cupboards.

When humans bait food with the poison, sodium fluorosilicate (, perennial grass is a more effective bait than growing lucerne, because of the dietary preferences of H. mossambicus.

In the Karoo, molehills ( and and and and and can easily be mistaken for the 'mounds' (soil dumps) of H. mossambicus.

My commentary:
This presumably refers to Cryptomys hottentotus.

Each colony of H. mossambicus forages over an area of up to 6.5 ha, which is far greater than that for Microhodotermes viator.

Coaton W G H (1962) Control of hodotermitid harvester termites in the Karoo. J. ent. Soc. S. Afr. 25(2): 318-327.

Microhodotermes viator is unusual among termites in making a 'hardcap' structure above ground. This is done in the Great Karoo, in a triangular part of its distribution with Beaufort West (, Aberdeen (,_South_Africa), and Willowmore ( at the points of the triangle.

The structure is a conical 'clay mound', about 1 m high, surrounded by a broad, bare, nearly flat apron of diameter about 6 m.

Posted by milewski 6 months ago

Add a Comment

Sign In or Sign Up to add comments