FILE: <bc-24.htm>                                                                                                                                                                                                              GENERAL INDEX          [Navigate to   MAIN MENU ]



                      CASTE BIAS Among Arthropods


---- Please CLICK on desired underlined categories [to search for Subject Matter, depress Ctrl/F ]:


Generalizations on Caste Determination





Historically castes have been recognized in arthropods for hundreds or perhaps thousands of years, but how they came about or were produced has remained a mystery. To this day there is still a lot of conjecture concerning the formation of castes.

Vandel (1930) stated that social insects are characterized by the existence of two kinds of females: reproductive and sterile (= queens and workers). The workers are neutral females that feed the larvae and perform all kinds of odd jobs around the colony. Haydak (1943) believed that in the honeybee the production of either a worker or a queen was due not to the change of food, but to the different amounts of essential nutrients consumed by the queen and worker larvae. He recognized that worker larvae decrease in weight after being sealed in their cells, while queen larvae actually grow after the sealing. He postulated that the anatomical and physiological differences between the worker and the queen honeybee are due partly to hormones, which are more activated in queens because of their superior nourishment. Worker larvae never get enough nourishment to produce the amount of hormone necessary for maturing the ovaries.

Light (1942) stated that the general problem of castes in social insects was the nature of the mechanisms which function in each generation of each species to cause the offspring of the same parents to develop, in relatively constant numbers, into several different types of individuals which possess the special morphological features and behavior patterns characteristic of the particular castes in the particular species, features which are correlated with the performance of species functions in the communal life. He believed that the food received by immature queens was different in nature to that received by immature worker bees. In ant colonies differential care and subsequent feeding were thought to produce the different castes. In termites, castes (involving males as well as females) were caused by extrinsic factors (actually intrinsic as we know it).

Schneirla & Brown (1952) observed that dry weather at the start of a brood has an impact upon the queen and colony of Eciton ants in Panama, which in some manner apparently effects the production of a temporary inhibition to fertilization and the production of a few queens and many males (no workers). Overfeeding, and effect of the workers eating most of the eggs forming the dry weather brood produce queens, so that the relatively few larvae are over-fed with booty plus eggs. The very few diploid eggs result from fertilization by residual sperm held over in queens' specially modified sperm duct after an all-worker brood.

Flanders (1953) believed that castes are limited in Hymenoptera probably entirely to species in which ovulation is externally induced. In such species ovisorption (egg resorption) is an effect of delayed ovulation. In many social species (if not most) caste is a result of undernourishment of the embryo through an extraction of nutriment from ripe eggs in the posterior (caudal) end of the ovariole. In species having many ovarioles, all ovulated eggs are worker-biased. Caste may be genetically limited to one sex in most Hymenoptera because it is based on imaginal (ovarian) diapause. Possibly in social species all castes are female because males are never undernourished during development. In such social Hymenoptera extra larval nutrition of workers counteracts its caste bias. Much is dependent on the response of the queen (female) to her environment.

Partial ovisorption seems an adequate explanation for workers caste determination in social Hymenoptera. It explains the deposition, by a single female, of eggs with different contents and volume (eggs of ants, etc.). Complete ovisorption accounts for some females becoming nongravid without egg deposition.

Partial ovisorption could result in an undernourished embryo. It can determine the embryo's course of development, and may result in the production of the worker caste. Wasps and bumblebees (more primitive Hymenoptera than Apidae) show gradations of castes. Partial ovisorption also explains the production on nonviable eggs.

Highly developed castes of ants are thought to be genetically determined, but realized only phenotypically. Factors regulating the amount of nutrient extracted from the ripe eggs are thought to be environmental. Involved are relative humidity affecting the ovisorption rate, and oviposition response of the female which regulate the amount of exposure of the eggs to the ovisorption process.

Prototypes of castes are found in ants. There are no structurally developed castes incapable of copulation in wasps and bumblebees. There are temporary gravid and nongravid types of females in some pteromalids (e.g., Peridesmia, Spintherus, Dibrachoides).

Wilson (1953) rejected Flanders' hypothesis of caste determination being based on ovisorption. He believed that the preponderance of available evidence indicates that the caste of female individuals is determined in the larval period, without regard to the original condition of the egg. He admitted that Flanders' idea involving ovisorption had a good chance of holding under conditions of complete dimorphism (e.g., honeybee); but for other species he thought that ovisorption might exercise a subsidiary influence in caste determination by statistically affecting the chance of a larva attaining the important size levels during its growth. In his final argument against Flanders' hypothesis, Wilson in effect embraced the hypothesis afterall! But only after a display of text sassyness and confusion! [see Wilson 1963, 1968].

In an elaboration on his hypothesis, Flanders (1957) cites evidence by Bier (1954) on Formica rufa showing that the worker derives only from ovarian eggs which have a reduced amount of yolk. The predisposition to become a worker, however, could be counteracted by heavy feeding during the early stages of larval development.

It is thought that queens of highly organized hymenopteran societies (e.g., army ant, honey bee) have lost the capacity to deposit yolk-replete eggs. The occasional occurrence of a worker-sized queen may be the result of an egg developing into reproductive females regardless of the nutrition of the larvae. All the fertilized eggs deposited by such a queen are predisposed, by the reduced amount of yolk, to become workers. The larvae from such eggs, however, may become queens if they receive materials that inhibit development into workers and stimulate development into queens. Flanders considered the assumption false that the worker caste in the honeybee is initiated during the larval stage.

Snodgrass (1956) referred to a larval diet containing specifically inhibitory ingredients that suppress the development of worker characters. The larger size of a queen cell also may inhibit the development of a worker therein. This comes about because all the food must be consumed or pupation cannot occur.

Flanders concluded that the queen in the more primitive social species may be derived either from a fertilized, yolk-replete egg or from an adult worker in which nutricial castration is psychologically counteracted. He believed that the ovigenic-ovisorptive cycle which characterizes the social Hymenoptera plays, in the honeybee, a basic role in the economy of the species by (1) allowing greater flexibility in oviposition so that the queen can take full advantage of the number of brood cells as they become available, and (2) the fact that the ovigenic-ovisorptive cycle correlates with numerous ovarioles to predispose all the eggs deposited by the queen to become worker-biased (males if unfertilized).

It appears that in the ant, caste formation is initiated in the ripe ovarian eggs. As Whiting (1938) suggested, slight stimuli applied at a nutritive-effective period in the ant egg could account for the wide differences between the worker, soldier and sexual female.

Generalizations on Caste Determination

 (S. E. Flanders, pers. commun. to E. F. Legner)

The worker caste consists of female adults in imaginal diapause. Imaginal diapause is facultative and reversible in parasitic species and in primitive social forms. Imaginal diapause occurs only in species in which reproduction is synovigenic, ovulation is externally induced, and unovulated eggs are resorbed (the number of eggs resorbed may exceed those deposited.

In the higher social Hymenoptera imaginal diapause is obligatory and is fixed by morphological differentiation. Some workers may generate eggs if queens are lacking, but these are able to produce worker progeny only by thelytoky.

Caste formation (imaginal diapause) is adaptive, being phenotypic in inception. A single queen when her environment is appropriate, can repeat a production sequence of workers, soldiers, males and queens. Colonies of ants living parasitically on other ant colonies may never produce "workers."

Imaginal diapause is initiated prior to ovulation. Normally, all ovarian eggs are male (haploid) and all are subject to possible resorption. Partially resorbed eggs are deposited without impairing their development. If too much yolk has been extracted the embryos die in all stages of development.

Adult progeny from eggs generated in one ovary of a single female may consist of both diapause and nondiapause females. Logically this is an effect of differential embryonic nutrition. Larval nutrition by counteracting the effects of embryonic malnutrition can change a prospective diapause female (worker) into a nondiapause "queen." However, the parasitic extraction of assimilated food from a prepupal queen ant can cause it to become a worker-like adult.


Exercise 24.1--Define caste bias.

Exercise 24.2--How many different castes are known among arthropods?

Exercise 24.3--How may castes be determined?



REFERENCES:        [Additional references may be found at  MELVYL Library ]


Bellows, T. S., Jr. & T. W. Fisher, (eds) 1999. Handbook of Biological Control: Principles and Applications. Academic Press,

        San Diego, CA.  1046 p.


Flanders, S. E. 1953. Caste determination in the social Hymenoptera. Sci. Mon. 76(3): 142-48.

Flanders, S. E. 1957. Regulation of caste in social Hymenoptera. J. New York. Ent. Soc. 65: 97-105.

Flanders, S. E. 1960. Caste in the honey bee. Insectes Sociaux 7: 7-16.

Haydak, M. H. 1943. Larval food and development of castes in the honeybee. J. Econ. Ent. 36: 778-92.

Light, S. F. 1942. The determination of the castes of social insects. Quart. Rev. Biol. 17: 312-06.

Light, S. F. 1943. The determination of the castes of social insects II. Quart Rev. Biol. 18: 46-63.

Schneirla, T. C. & R. Z. Brown. 1952. Sexual broods and the production of young queens in two species of army ants.

       Zoologica 37: 5-32.


Vandel, A. 1930. La production d' intercastes chez la fourmi Pheidole pallidula sous l'action de parasites du genre Mermis.

       Bull. Biol. France & Belg. 64: 457-92.


Wilson, E. O. 1953. On Flanders' hypothesis of caste determination in ants. Psyche 60: 15-20.

Wilson, E. O. 1963. The social biology of ants. Ann. Rev. Ent. 8: 345-68.

Wilson, E. O. 1968. The ergonomic of castes in the social insect. Amer. Nat. 102(923): 41-6.