The methods described here do not all have the same level of evidence. Certain early interventions, notably the use of young queens, have direct experimental support. Others — Demaree, reversing brood boxes, VIRDIS or temporary caging of the queen — rely mainly on long beekeeping experience and biologically plausible mechanisms. The absence of a modern comparative trial is not evidence of ineffectiveness; above all, it prevents the effect from being quantified precisely, the stage at which it is most effective from being identified, and the mechanism from being isolated.

1. When does swarm preparation actually begin?

Brood and population development varies with the year, climate, nectar flow and management system; there is no universal date for the spring maximum.

Swarm preparation begins before the beekeeper discovers the first occupied queen cells. The colony gradually shifts from a state dominated by growth towards a reproductive state. Several dimensions change together: adult population, brood demography, occupied space, food stores, comb-building activity, the physiology of the queen and workers, and social interaction networks.

No universal trigger has been demonstrated. A colony may change over several days or weeks before queen rearing becomes visible, and individual stages of the process may still be abandoned. It is therefore more accurate to speak of a progressive change in colony state than of a switch that suddenly changes the colony from “not preparing to swarm” to “preparing to swarm”.

The older explanation according to which the process begins when the queen’s movements become restricted and a pheromone produced by her legs is no longer distributed sufficiently is too simple. Information originating from the queen does not depend on a single “marking” of the combs and does not behave like a homogeneous cloud whose concentration automatically decreases as the colony grows. Queen fertility signals are partly distributed through contact networks among workers (Richardson et al., 2024). An earlier experiment also showed that a stationary source of synthetic queen mandibular pheromone became less effective in strongly congested colonies, whereas broader distribution retained more of its effect (Winston et al., 1991). This finding is compatible with an influence of spatial organisation on queen communication, but does not demonstrate that congestion alone “switches off” a pheromonal network.

The queen is also not the only source of reproductive information. Queen mandibular pheromone can inhibit the rearing of new queens, but its effect decreases when it acts alone for several days (Pettis et al., 1995). Eggs and very young larvae provide additional information about the queen’s effective reproductive function (Pettis et al., 1997). These experiments mainly concern regulation of queen rearing after experimental manipulation; they do not demonstrate that a decline in one of these signals triggers natural swarming.

The classic experiments show above all that several phenomena commonly grouped under the term “lack of space” must be distinguished. Restricting the space available to adult bees favoured swarming, whereas restricting the cells available for egg laying did not produce the same effect (Simpson & Riedel, 1963). Severe restriction of laying area alone likewise did not produce the marked increase in queen rearing that would be expected from a simple “not enough room for the queen to lay” explanation (Simpson & Greenwood, 1975).

These findings do not mean that space is unimportant. Small colonies allowed to continue growing in small hives until they actually exceeded their available capacity developed occupied queen cells much more frequently and often swarmed (Simpson & Moxley, 1971). Restricting total hive volume could also increase queen rearing (Simpson, 1973), while increasing volume and ventilation reduced the conversion of queen cups into queen cells and reduced swarm departures under the conditions studied by Lensky and Seifert (1980). By contrast, suddenly reducing the volume available to already developed colonies did not necessarily increase swarming (Caron, 1981).

The conclusion is therefore more precise than “more space = less swarming”: the outcome depends on what is actually restricted, when the restriction occurs and the state the colony is already in.

In practice, brood-nest congestion brings several of these dimensions together. It may include a large adult population, extensive capped brood close to emergence, relatively little young brood, fewer immediately usable cells, nectar or pollen stored within the brood nest and changes in spatial organisation. It is therefore neither a single mechanism nor a certain diagnostic test, but an integrated warning sign.

Queen cells should likewise be interpreted as markers of stage. An empty queen cup is only weakly informative. A cell containing an egg represents a further step, but queen rearing may still be abandoned. A young larva being actively fed in a swarm cell shows that the colony is now genuinely investing in queen rearing. Several capped swarm cells together with other converging signs indicate an advanced stage in practical terms. None of these observations, however, constitutes a universal point of no return when considered in isolation (Allen, 1965).

2. Which factors favour swarming?

Treat risk factors as a dynamic configuration rather than as a list of independent triggers.

No single factor makes it possible to predict a swarm departure with certainty. The following factors can, however, alter swarming probability or help the beekeeper recognise a high-risk configuration:

• Genetics: swarming propensity has a genetic component and can respond to selection. European comparisons nevertheless show a strong interaction between genetic origin and environment; a single swarming event is therefore not a genetic test (Uzunov et al., 2014).
• Queen age: colonies headed by young queens generally show less swarm preparation. The trials of Forster (1969) and Hauser and Lensky (1994) support this effect. A young queen reduces probability; she does not make swarming impossible.
• High density of adult bees: this is among the best-supported early factors experimentally (Simpson & Riedel, 1963).
• Large emergence waves: extensive areas of capped brood can rapidly increase the adult population. This is not a swarming threshold, but it is an important demographic component.
• Brood-nest congestion and use of space: nectar storage, comb building, occupation by bees, forthcoming emergence and egg laying use or modify the same comb areas. Functional space and the overall colony state therefore matter more than laying area alone.
• Season and nectar flow: in temperate regions, reproductive swarming generally concentrates during the period of rapid spring growth, but the window varies with altitude, climate, phenology and year.
• Weather: several unfavourable days may mainly delay the departure of colonies already prepared to swarm and then produce a wave of swarms once flying conditions improve (Henneken et al., 2012). The complete causal chain “poor weather → increased bee density → triggering of swarming” remains plausible but unproven.
• Heavy liquid feeding during rapid colony growth: this can alter storage and cell use. Syrup can be moved and redistributed within the nest (Free & Spencer-Booth, 1961; Eyer et al., 2016; Colin et al., 2018). A direct causal chain “syrup → swarming” has not been demonstrated.
• Volume, heat and ventilation: these factors can alter queen rearing and colony state in certain configurations, but neither high temperature nor low volume is by itself a universal trigger (Lensky & Seifert, 1980; Caron, 1981).

3. Prevent, modify or control the loss?

Choose the method according to the biological stage and the actual objective of the intervention.

Not all measures traditionally described as “swarm prevention” pursue the same objective. Three situations should be distinguished.

• Prevent the establishment of stable swarm preparation: adapt functional space early enough, add supers at the appropriate time, use selected young queens and, depending on the management system, remove bees or brood to form nucleus colonies before queen rearing has become durably established.
• Modify preparation that is already under way: once larvae are being actively fed in swarm cells, the objective is no longer merely to reduce a risk factor. Removing or redistributing bees and brood, controlled division, Demaree, VIRDIS or queen caging may alter several components of colony state. For most of these methods, however, a biological “reset” of advanced swarm preparation has not been experimentally demonstrated.
• Control loss without eliminating the swarming state: when departure is close, certain measures mainly prevent the loss of the queen or swarm. Preventing the loss of a swarm is not the same as preventing swarming.

3.1 Simple preventive measures

Providing genuinely usable space early enough, adding supers at the appropriate time and adapting brood-box volume to colony strength are among current Swiss practical recommendations (Bee Health Service [BHS], 2026). These measures are biologically consistent with findings on density, volume and congestion, but their independent effect on natural swarm departure remains insufficiently quantified. In particular, they should not be reduced to “giving the queen more room to lay”.

The use of young queens has more direct experimental support. Forster (1969) observed a strong reduction in swarm preparation with queens reared in spring, while Hauser and Lensky (1994) observed fewer swarm cells in colonies with younger queens.

Providing drone comb may form part of spring colony management. By contrast, drone brood removal is not a demonstrated swarm-prevention method. In a controlled trial on A. m. ligustica, removing drone pupae did not significantly alter swarming behaviour (Zheng Zhijiang et al., 2000).

Shading or good thermal management may be useful in very hot situations, but should not be presented as universal methods of swarm prevention.

3.2 Removing queen cells

Removing queen cells prevents the development of the queens concerned, but it does not demonstrate that the colony’s reproductive state has disappeared. As long as several components of preparation persist, new cells may be constructed.

The developmental stage of the removed cells also matters. Removing cells containing only a newly laid egg is biologically different from removing several capped swarm cells from a colony in which the old queen has already substantially reduced laying and other signs indicate that departure is close.

There is no general evidence that removing queen cells “accelerates” swarming. For Swiss practice, the BHS currently recommends, when this strategy is chosen, removing all swarm cells every 7 to 9 days (BHS, 2026). This interval is a management recommendation, not evidence for a biological point of no return occurring on a specific day.

3.3 Removing bees or brood

“Removing brood” may refer to biologically very different interventions. Removing capped brood with almost no bees mainly reduces the future emergence wave. Removing brood together with its bees reduces both the present and future population. Removing mainly adult bees acts immediately on density. Merely redistributing brood within the same colony primarily changes spatial organisation.

These differences explain why there is no single general effect of “brood removal”. Radtke (2010) tested different forms of capped-brood removal in A. m. carnica colonies without finding a uniform protective effect against swarming indicators. Removing capped brood alone therefore does not constitute a universal reset.

By contrast, Maucourt et al. (2018) observed less swarming in donor colonies after nucleus colonies had been formed. This intervention genuinely removed part of the colony’s demography, but the component responsible for the effect was not isolated.

The useful practical question is therefore: what is actually being removed — present bees, future bees, or both?

3.4 Queen clipping

Clipping part of one wing can prevent the old queen from following the prime swarm normally. Simpson (1963) showed that a swarm deprived of a queen capable of flight can return towards its colony of origin. The biological swarm preparation may nevertheless persist.

Queen clipping is therefore primarily a form of loss control. It may provide the beekeeper with additional time or make recovery of the queen easier, but it does not demonstrate that the swarming state has been eliminated. Forster (1971) also provides data on the technical tolerance of clipping under the conditions studied, not on its ability to eliminate the reproductive state.

4. Method 1: the artificial swarm

Anticipate or channel natural colony division by deliberately creating a new broodless unit.

The artificial swarm is a form of controlled division. At an advanced stage of swarm preparation, this approach may be biologically more coherent than attempting at all costs to restore the colony’s original state: the beekeeper anticipates part of the natural division and retains control over location, queen and subsequent management of the resulting units.

It is not simply a matter of “making the colony believe” that it has swarmed. A new unit is genuinely created from a queen and a mass of workers. Depending on the variant used, this may also profoundly alter the population and organisation of the donor colony.

The procedures vary according to whether the old queen or a new selected queen is used, and whether the new unit remains at the same apiary or is moved to another location.

One possible practical sequence is as follows:

1. Choose the queen. Use the donor colony’s old queen or a new selected queen according to the management objective.
2. Form the population. Take a sufficient mass of bees and place them in a well-ventilated swarm box. The quantity depends on hive format, season and objectives; traditional figures of 1 to 2 kg are practical benchmarks rather than a biological constant.
3. Introduce the queen. If a new queen is introduced in a cage, use the appropriate introduction procedure.
4. Optional cellar confinement. In traditional variants that include it, keep the artificial swarm cool, dark and well ventilated. The duration depends on the method, temperature, unit size and future location.
5. Hive the swarm. Install it on foundation or other surfaces suitable for comb building.
6. Feed as required. Feeding depends on nectar flow, available stores and the size of the unit. Systematically heavy feeding is unnecessary during a good nectar flow.
7. Use the broodless window. The broodless phase can be incorporated into the Varroa management concept in accordance with current Swiss recommendations (BHS, 2025).
8. Inspect the new unit. Check queen acceptance and activity, comb building, stores and development.
9. Inspect the donor colony as well. Creating a new unit does not automatically demonstrate that all swarm preparation has disappeared from the colony of origin.

See also: Practical Guide: 1.4.2 Artificial swarm and Practical Guide: 1.4.3 Artificial swarm with queen.

5. Method 2: reversing the brood boxes

Reversing brood boxes is a practice used mainly in stack hives wintered on two boxes. After winter, the cluster and first brood may be located mainly in the upper box, while a substantial part of the lower box has become free.

Reversing the two boxes at the appropriate time places part of the drawn comb space above the brood again and alters vertical movement within the colony. A further reversal may sometimes be made later once the brood in the other box has emerged.

This practice may:

• alter the spatial distribution of brood and bees;
• make drawn cells available for expansion of the brood nest;
• alter storage zones and vertical movement;
• change local densities and bee traffic.

It would, however, be too strong to conclude that reversing boxes reliably prevents swarming or that one specific mechanism explains any effect it may have. Its use relies largely on experience with stack-hive systems, while its independent effect on natural swarm departure has rarely been isolated experimentally.

See also: Understanding swarming.

6. Method 3: Demaree

The Demaree method is a long-established swarm-control practice that seeks to retain a large part of the colony’s productive strength while spatially separating the queen from much of the brood.

Several variants exist. In a classic arrangement, the queen is kept in the lower box with one brood frame and several drawn frames or free comb areas. Most of the brood is moved to an upper box separated from the queen by a queen excluder and often by the honey supers.

This reorganisation simultaneously alters:

• the position of the brood;
• the distribution of nurse bees;
• local densities around the queen and brood;
• the areas available for egg laying;
• bee traffic between different parts of the hive;
• contact networks between queen, brood and workers;
• the future location of food storage.

Emergency or other queen cells may be reared in the upper part when the bees there receive insufficient functional information from the queen. Inspecting and managing these cells therefore forms an important part of many variants.

Two overly simple traditional explanations are best avoided:

• the bees in the upper box are not simply unable to “sense” the queen; queen-derived information depends on contact networks and several reproductive signals;
• the colony does not literally “believe” that it has swarmed. It has undergone a profound spatial, demographic and social reorganisation.

The Demaree method has a long beekeeping tradition. Modern literature, however, does not yet provide a robust estimate of its specific efficacy stratified according to the exact stage of natural swarm preparation. It is therefore reasonable to present it as biologically plausible and widely proven in practice, but not as an experimentally demonstrated reset of advanced swarm preparation.

7. Method 4: VIRDIS

The VIRDIS method follows a different logic from classic demographic removal: capped brood is not removed from the colony, but moved into the honey supers. The workers that subsequently emerge therefore remain part of the production colony (Colombari, 2019).

1. When the colony strongly occupies the brood box and has extensive brood, add the queen excluder and the supers required by the system.
2. Remove two or three frames consisting mainly of capped brood from the brood box and replace them with drawn combs or frames to be built. The queen must remain in the lower brood box.

Place the capped-brood frames in the centre of the supers. Adjust the remaining space to the colony’s actual strength and, if necessary, use temporary division boards to avoid excessive volume.

4. Once the workers have emerged from the transferred frames, these frames can be moved back down and new frames of capped brood moved upwards.
5. This rotation may be repeated during the growth phase according to colony strength, weather and nectar flow.
6. When the nectar flow becomes strong, management is adapted to honey-storage requirements and the production objective.

The method therefore repeatedly alters brood position, the distribution of nurse bees, laying and comb-building areas, local densities, bee traffic and use of hive volume. Unlike classic brood removal, it specifically seeks to retain the future production population.

It would, however, be too strong to claim that maintaining a “stable ratio between open and capped brood” by itself prevents queen-cell construction. No such threshold has been demonstrated.

Particular vigilance is required under Swiss spring conditions. VIRDIS can rapidly increase the available hive volume. During a prolonged return of cold weather, the upper part may become much less densely occupied and functional access to food stores may deteriorate. Close monitoring is therefore advisable and, if necessary, volume can be reduced temporarily, for example by removing a lightly occupied super for a period. This recommendation is based on practical management experience rather than an experimentally established threshold.

8. Method 5: temporary caging of the queen

The Sion beekeeping section has developed practical experience with temporary caging of the queen in a Scalvini-type isolation cage. The objective is to intervene in a colony in which swarm preparation is already clearly under way while keeping the queen and most of the population in the same hive.

Cage the queen. Install the cage in a drawn comb within the brood nest, in an area well occupied by bees. Maintain sufficient storage space. During a nectar flow, prevent cells gradually becoming free in the brood nest from being immediately filled with nectar. Remove all existing swarm cells. Inspect again after about five days and remove any newly constructed queen cells. This interval belongs to the ApiSion practical protocol and is not a universal biological rule. From about fourteen days onwards, reassess the colony. Examine brood structure, queen cells, genuinely usable space, storage patterns and the overall colony state rather than relying on a fixed numerical ratio between open and capped brood. Release the queen when conditions are judged favourable and then check the resumption of laying and the absence of renewed active queen rearing.

Caging simultaneously alters egg laying, the production of very young brood, future emergence waves, cell use and the reproductive physiology of the queen. Aamidor et al. (2022) show that queen ovarian function is plastic, while Kovačić et al. (2023), in a different context, show that prolonged caging can create a controlled brood interruption whose consequences depend on the timing of the intervention.

These studies do not, however, directly test the ApiSion protocol as a treatment for natural swarm preparation already under way. They provide a physiological basis for its plausibility, but do not demonstrate that a Scalvini cage alone “switches off” the swarming state.

Over the following days, the protocol can alter several components of brood-nest congestion: existing brood ages and emerges, production of new brood is strongly reduced, comb areas become free, the next emergence wave changes and visible queen rearing is interrupted by removal of the queen cells. It is therefore biologically plausible that the colony’s overall trajectory is altered.

See also: A new method for extinguishing the swarming mood.

There is therefore no universally “best” method for controlling swarming. The earlier the intervention, the more likely it is to act on the conditions that make the reproductive transition probable before that state becomes established. The more advanced the preparation, the more an intervention must take account of the fact that the colony has already changed its demography, the physiology of its queen and some of its workers, its resources and its social organisation.

At an advanced stage, deliberately controlling the division may sometimes be biologically more coherent than attempting to achieve a “reset” whose existence has not been demonstrated.

---

See also:

• Swarming in practice: preventing, controlling and making use of it
• A new method for extinguishing the swarming mood
• Practical Guide: 4.7.5. Managing the swarming mood
• Swarming: recognising the stage and acting at the right moment
• Queen cells
• Making increase
• Varroa: brood interruption

References

• Aamidor, S. E., Cardoso-Júnior, C. A. M., Harianto, J., Nowell, C. J., Cole, L., Oldroyd, B. P., & Ronai, I. (2022). Reproductive plasticity and oogenesis in the queen honey bee (Apis mellifera). Journal of Insect Physiology, 136, 104347. https://doi.org/10.1016/j.jinsphys.2021.104347
• Allen, M. D. (1965). The production of queen cups and queen cells in relation to the general development of honeybee colonies, and its connection with swarming and supersedure. Journal of Apicultural Research, 4, 121–141. https://doi.org/10.1080/00218839.1965.11100115
• Caron, D. M. (1981). Congestion, seasonal cycle and queen rearing as they relate to swarming in Apis mellifera L. Annals of the Entomological Society of America, 74(1), 134–137. https://doi.org/10.1093/aesa/74.1.134
• Colin, T., Bruce, J., Meikle, W. G., & Barron, A. B. (2018). The development of honey bee colonies assessed using a new semi-automated brood counting method: CombCount. PLOS ONE, 13, e0205816. https://doi.org/10.1371/journal.pone.0205816
• Colombari, L. (2019). Gestione della sciamatura: metodo “Virdis”. L’Apis, 4/2019.
• Eyer, M., Neumann, P., & Dietemann, V. (2016). A look into the cell: Honey storage in honey bees, Apis mellifera. PLOS ONE, 11, e0161059. https://doi.org/10.1371/journal.pone.0161059
• Forster, I. W. (1969). Swarm control in honey bee colonies. New Zealand Journal of Agricultural Research, 12, 605–610. https://doi.org/10.1080/00288233.1969.10421245
• Forster, I. W. (1971). Effect of clipping queen honey bees' wings. New Zealand Journal of Agricultural Research, 14, 535–537. https://doi.org/10.1080/00288233.1971.10427116
• Free, J. B., & Spencer-Booth, Y. (1961). The effect of feeding sugar syrup to honey-bee colonies. The Journal of Agricultural Science, 57, 147–151. https://doi.org/10.1017/S0021859600047614
• Hauser, H., & Lensky, Y. (1994). The effect of the age of the honey bee queen on worker population, swarming and honey yields in a subtropical climate. Apidologie, 25, 566–578. https://doi.org/10.1051/apido:19940607
• Henneken, R., Helm, S., & Menzel, A. (2012). Meteorological influences on swarm emergence in honey bees as detected by crowdsourcing. Environmental Entomology, 41(6), 1462–1465. https://doi.org/10.1603/EN12139
• Kovačić, M., Uzunov, A., Tlak Gajger, I., Pietropaoli, M., Soroker, V., Adjlane, N., Benko, V., Charistos, L., Dall'Olio, R., Formato, G., Hatjina, F., Malagnini, V., Freda, F., Otmi, A., Puškadija, Z., Villar, C., & Büchler, R. (2023). Honey vs. mite—A trade-off strategy by applying summer brood interruption for Varroa destructor control in the Mediterranean region. Insects, 14(9), 751. https://doi.org/10.3390/insects14090751
• Lensky, Y., & Seifert, H. (1980). The effect of volume, ventilation and overheating of bee colonies on the construction of swarming queen cups and cells. Comparative Biochemistry and Physiology Part A: Physiology, 67, 97–101. https://doi.org/10.1016/0300-9629(80)90413-2
• Maucourt, S., Fournier, V., & Giovenazzo, P. (2018). Comparison of three methods to multiply honey bee (Apis mellifera) colonies. Apidologie, 49(3), 314–324. https://doi.org/10.1007/s13592-017-0556-9
• Pettis, J. S., Winston, M. L., & Collins, A. M. (1995). Suppression of queen rearing in European and Africanized honey bees Apis mellifera L. by synthetic queen mandibular gland pheromone. Insectes Sociaux, 42(2), 113–121. https://doi.org/10.1007/BF01242447
• Pettis, J. S., Higo, H. A., Pankiw, T., & Winston, M. L. (1997). Queen rearing suppression in the honey bee: Evidence for a fecundity signal. Insectes Sociaux, 44(4), 311–322. https://doi.org/10.1007/s000400050053
• Radtke, J. (2010). Einfluss der Brutentnahme bei der Honigbiene Apis mellifera auf die Leistung der Völker und ihre Parasitierung mit Varroa destructor [Dissertation, Martin-Luther-Universität Halle-Wittenberg]. https://doi.org/10.25673/455
• Richardson, T. O., Kay, T., Keller, L., & Stroeymeyt, N. (2024). Pheromone relay networks in the honeybee: Messenger workers distribute the queen's fertility signal throughout the hive. BMC Biology, 22, 288. https://doi.org/10.1186/s12915-024-02083-w
• Bienengesundheitsdienst (BGD). (2025). Übersicht Methoden zur Jungvolkbildung (Merkblatt 1.4, V 2511).
• Bienengesundheitsdienst (BGD). (2026). Schwarmstimmung: Was tun?
• Simpson, J. (1963). Queen perception by honey bee swarms. Nature, 199, 94–95. https://doi.org/10.1038/199094a0
• Simpson, J. (1973). Influence of hive-space restriction on the tendency of honeybee colonies to rear queens. Journal of Apicultural Research, 12(3), 183–186. https://doi.org/10.1080/00218839.1973.11099747
• Simpson, J., & Greenwood, S. P. (1975). Results of restricting the brood space of honeybee colonies. Journal of Apicultural Research, 14(1), 51–55. https://doi.org/10.1080/00218839.1975.11099801
• Simpson, J., & Moxley, E. (1971). The swarming behaviour of honeybee colonies kept in small hives and allowed to outgrow them. Journal of Apicultural Research, 10(3), 109–113. https://doi.org/10.1080/00218839.1971.11099681
• Simpson, J., & Riedel, I. B. M. (1963). The factor that causes swarming by honeybee colonies in small hives. Journal of Apicultural Research, 2(1), 50–54. https://doi.org/10.1080/00218839.1963.11100056
• Uzunov, A., Costa, C., Panasiuk, B., Meixner, M. D., Kryger, P., Hatjina, F., Bouga, M., Andonov, S., Bieńkowska, M., Le Conte, Y., Wilde, J., Gerula, D., Kiprijanovska, H., Filipi, J., Petrov, P., Ruottinen, L., Pechhacker, H., Berg, S., Dyrba, W., Ivanova, E., & Büchler, R. (2014). Swarming, defensive and hygienic behaviour in honey bee colonies of different genetic origin in a pan-European experiment. Journal of Apicultural Research, 53(2), 248–260. https://doi.org/10.3896/IBRA.1.53.2.06
• Winston, M. L., Higo, H. A., Colley, S. J., Pankiw, T., & Slessor, K. N. (1991). The role of queen mandibular pheromone and colony congestion in honey bee reproductive swarming. Journal of Insect Behavior, 4(5), 649–660. https://doi.org/10.1007/BF01048076
• Zheng Zhijiang, Wu Guisheng, Zhang Zhongyin, & Gu Dongsheng. (2000). Effects of drone pupae excised on the colony productivity, reproductive capacity and swarming. Journal of Zhejiang University Agriculture and Life Sciences, 26, 540–542. https://doi.org/10.3785/1008-9209.2000.05.0540