animals they copy, andsubsequently evaluate theusefulness of the copiedbehaviour, or the usefulness ofthe particular model in general.The study of heterospecificinformation transfer could thus bea useful avenue of futureresearch, in both insects and theless successful other animals thatpopulate the planet.

References1. Galef, B.G., and Laland, K.N. (2005).

Social learning in animals: Empiricalstudies and theoretical models.Bioscience 55, 489–499.

2. Danchin, E., Giraldeau, L.A., Valone, T.J.,and Wagner, R.H. (2004). Publicinformation: From nosy neighbors tocultural evolution. Science 305, 487–491.

3. Romanes, G.J. (1884). Mental evolutionin animals (New York: AMS Press).

4. Galef, B.G. (1996). Introduction. In Sociallearning in animals, C.M. Heynes andB.G. Galef, eds. (San Diego: AcademicPress).

5. Hopkins, A.D. (1917). A discussion ofC.G.Hewitt’s paper on ‘Insect Behavior’.J. Econ. Entomol. 10, 92–93.

6. Barron, A.B. (2001). The life and death ofHopkins’ host-selection principle. J.

Insect Behav. 14, 725–737.7. Technau, G., and Heisenberg, M. (1982).

Neural reorganization duringmetamorphosis of the corporapedunculata in Drosophila melanogaster.Nature 295, 405–407.

8. Tully, T., Cambiazo, V., and Kruse, L.(1994). Memory through metamorphosisin normal and mutant Drosophila. J.Neurosci. 14, 68–74.

9. Williams, N.M. (2003). Use of novel pollenspecies by specialist and generalistsolitary bees (Hymenoptera:Megachilidae). Oecologia 134, 228–237.

10. Kirchner, W.H. (1987). Tradition imBienenstaat. Kommunikation zwischenImagines und der Brut der Honigbienedurch Vibrationssignale. PhD Thesis,Wuerzburg University, supervised by M.Lindauer.

11. v Frisch, K. (1967). The dance languageand orientation of bees (Cambridge:Harvard Univ. Press).

12. Farina, W.M., Grueter, C., and Diaz, P.C.(2005). Social learning of floral odoursinside the honeybee hive. Proc. R. Soc.Lond. B. Biol. Sci. 272, 1923–1928.

13. D’Adamo, P., and Lozada, M. (2005).Conspecific and food attraction in thewasp Vespula germanica (Hymenoptera:Vespidae), and their possiblecontributions to control. Ann. Entomol.Soc. Am. 98, 236–240.

14. Leadbeater, E., and Chittka, L. (2005). Anew mode of information transfer inforaging bumblebees? Curr. Biol. 15,R447–R448.

15. Worden, B.D., and Papaj, D.R. (2005).Flower choice copying in bumblebees.Biology Letters DOI:10.1098/rsbl.2005.0368.

16. Chittka, L., Thomson, J.D., and Waser,N.M. (1999). Flower constancy, insectpsychology, and plant evolution.Naturwiss. 86, 361–377.

17. Coolen, I., Dangles, O., and Casas, J.(2005). Social learning in non-colonialinsects? Curr. Biol. 15, this issue.

18. Parejo, D., Danchin, E., and Aviles, J.M.(2005). The heterospecific habitatcopying hypothesis: can competitorsindicate habitat quality? Behav. Ecol. 16,96–105.

19. Rainey, H.J., Zuberbuhler, K., and Slater,P.J.B. (2004). The responses of black-casqued hornbills to predatorvocalisations and primate alarm calls.Behaviour 141, 1263–1277.

20. Nieh, J.C., Barreto, L.S., Contrera, F.A.L.,and Imperatriz-Fonseca, V.L. (2004).Olfactory eavesdropping by acompetitively foraging stingless bee,Trigona spinipes. Proc. R. Soc. Lond. B.Biol. Sci. 271, 1633–1640.

School of Biological and ChemicalSciences, Queen Mary University ofLondon, Mile End Road, London E14NS, UK.

DOI: 10.1016/j.cub.2005.10.018

Dispatch R871

Patricia Wadsworth

When a eukaryotic cell divides,within minutes of anaphasechromosome motion the corticalcytoplasm begins to ingress at alocation overlying the positionpreviously occupied by thechromosomes at metaphase. Theremarkable ability of cells tospecify the site of contractile ringformation so precisely hasfascinated and frustratedbiologists for decades. New work[1–3] has now shown that activeRhoA forms a narrow zone at thesite where the contractile ring willform, and identified the RhoGTPase-activating protein(RhoGAP) component of thecentralspindlin complex and theGTP exchange factor for RhoA askey players in the activation ofRhoA.

Microtubules Specify the Site ofContractile Ring FormationIt has long been recognized thatsome component(s) of the mitoticspindle plays a key role indetermining the site of cleavagefurrow formation when aeukaryotic cell divides. Supportfor the idea that the spindledelivers a signal to the cortex hascome from experiments in whichthe spindle was repeatedlyrepositioned in an artificiallyelongated embryonic echinodermcell [4]. The results showed thatmultiple furrows can besequentially specified,demonstrating that the cortex ofthe anaphase cell is globallycompetent to furrow, providedthat the appropriate signal isdelivered and received.Micromanipulation experiments,also performed in echinoderm

blastomeres, showed that twoastral arrays of microtubules,lacking intervening chromosomes,are sufficient to generate thesignal for furrowing [4].

Subsequent work inmammalian and Drosophila cells,in which the geometry of spindle,asters and cortex differs fromthat in large, spherical embryoniccells, suggested that interzonal,not astral, microtubules arerequired for cytokinesis [5]. Giventhese conflicting results, mucheffort has been focused ondetermining which class, orclasses, of microtubules areresponsible for furrow induction.It is now generally agreed thatmicrotubules are the onlystructural component needed forfurrow induction [6], and that theclass, or classes, of microtubulesthat are required depends on celltype. In some cases, twosequential signals from astraland interzonal microtubules areused [5,7]. The finding thatdifferent arrangements ofmicrotubules contribute tospecification of furrowing indifferent organisms and thatmultiple signals may participate

Cytokinesis: Rho Marks the Spot

During cytokinesis of a eukaryotic cell, following the chromosomemovements of anaphase, a contractile ring forms in the cortex midwaybetween the segregating chromosomes and divides the cell into twodaughters. Recent studies have provided new insights into themechanism by which the site of contractile ring assembly is specified.

is consistent with the essentialnature of cytokinesis.

What Is the Nature of the Signalthat Microtubules Deliver to theCortex?In an anaphase cell, moleculesthat localize to astral and/orinterzonal microtubules, or to thecell cortex, could be part of thesignal or its delivery system.Genetic approaches haveidentified several conservedprotein complexes that localizeto interzonal microtubules in adividing cell. For example, thechromosomal passengercomplex — composed ofINCENP, Aurora B kinase,Borealin and Survivin —relocalizes from centromericchromatin to interzonalmicrotubules following anaphaseonset, and at least somecomponents of the complex havebeen shown to be required forcytokinesis [8].

Another protein complex,centralspindlin, localizes tointerzonal microtubules in aregion that likely corresponds tothe region of overlap betweenmicrotubules from the two half-spindles [5]. Centralspindlin iscomposed of a kinesin-relatedprotein, MKLP1, in mammaliancells (hereafter kinesin-6) and aGTPase-activating protein forRho, MgcRacGAP in mammals(hereafter GAP). The kinesin-6component of centralspindlincould participate in the deliveryof signals to the cortex bymoving cargo toward themicrotubule plus-ends. In thenematode Caenorhabditiselegans, localization ofcentralspindlin requires INCENP,and the kinesin-6 and aurora B

orthologs interact genetically,demonstrating that thesecomplexes, and possibly otherproteins that localize to thecentral spindle, may function inan integrated manner [5,8].

RhoGTPases, which regulatethe organization of the actincytoskeleton and myosin IIactivity, have long beensuspected to contribute to theregulation of cytokinesis [1].Rho’s activity is regulated by aguanine nucleotide exchangefactor, or RhoGEF (ECT2 inmammals, hereafter RhoGEF).Both RhoA and RhoGEFs arerequired for cytokinesis in diversesystems [1,9]. The DrosophilaRhoGEF, Pebble, has beenshown to interact with the GAPcomponent of centralspindlin,leading to a model in whichcentralspindlin complexes travelalong microtubules to the cortexwhere they associate with theRhoGEF and regulate formationof the contractile ring [10].

Analysing Cytokinesis with GFP-Fusion ProteinsTo determine if active RhoAcontributes to specification of thesite of furrow formation, Bementand co-workers [1] linked GFP toa fragment of rhotekin that bindsspecifically to active RhoA (RhoA-GTP). When this construct isexpressed in echinoderm orvertebrate embryonic cells,narrow zones of GFPfluorescence are observed in thecortex prior to and duringfurrowing. The width of the zoneremains constant duringcleavage. RhoA zones were alsodetected during the highlyasymmetric division thataccompanies polar body

extrusion, indicating that they area conserved feature of diversedivisions. When the spindle isrepositioned to one end of thecell with a blunt needle — amanipulation that is possibleusing these large pliableembryonic cells — the RhoAsignal disappears from theoriginal site and reappears at thenew location of the spindlemidzone. Active RhoA is notdetected when microtubules aredisrupted with nocodazole, but isunaffected following disassemblyof F-actin with latrunculin A. Thepharmacological andmicromanipulation results areconsistent with microtubulesdirecting the formation of adynamic zone of active RhoA [1].

What Is the Molecular Basis forthe Local Activation of RhoA?To examine this question, Yuce etal. [3] used RNA interference(RNAi) to knockdown expressionof candidate molecules in humancells and imaged living cellsexpressing YFP-RhoA. They foundthat depletion of the central-spindle-associated RhoGEF,ECT2, completely eliminates thezone of active RhoA and preventsboth accumulation of myosin IIand actin in the contractile ringand ingression [3]. Because theRhoGEF and GAP component ofcentralspindlin have been shownto interact [10], the authorsexamined the phenotype of cellslacking the GAP, MgcRacGAP.Depletion of the GAP blockedRhoA activation and contractilering assembly and ECT2 was nolonger localized to the disruptedcentral spindle microtubules.

Depletion of the kinesin-6subunit of centralspindlin also

Current Biology Vol 15 No 21R872

Figure 1. Rho in cytokinesis.

(A) Molecular interactions required for activation of RhoA. Only Rho molecules near microtubules receive the signal to activate Rho.(B) Steps leading to furrow specification: (i) Kinesin-6 transports signaling complexes (arrowheads) outward along dynamicmicrotubules (red). (ii) RhoA is locally activated (yellow boxes) at sites where signaling is concentrated, leading to furrowing. (iii) RhoAstabilization of microtubules (curved arrows) could promote continued delivery.

VV

V

V

VVV

V

V

V

V

(i) (ii) (iii)

V

VV

V

V

V

V

V

V

V

VVV

V

VV

V

V

V

V

Kinesin6/GAPR

R*

RhoGEFMT

R

R

R Rho

A B

Current Biology

disrupted interzonal microtubulesand ECT2 localization. In thesecells, however, diffuse zones ofRhoA were detected and thedegree to which the RhoA wasfocused correlated with thecapacity for ingression. Thisexperiment shows that thelocalization of the RhoGEF to thecentral spindle is not strictlyrequired for furrowing, but isrequired to generate a discrete,narrow zone of active RhoA. Howthe RhoA becomes localized incells with diffuse GEF is not yetknown. These results alsoindicate that the GAP mustcontribute to RhoA zoneformation by a mechanismdistinct from central spindleformation. Independentexperiments using RNAi todeplete GAP or RhoGEF confirmthat contractile ring proteins,including annillin andphosphorylated myosin lightchain, fail to localize to the cortexin depleted cells [2].

To understand how the GAPmight regulate RhoGEF duringcytokinesis, biochemicalapproaches were used to examinepotential interactions. RhoGEFand GAP interact in a cell cycle-dependent manner, with maximalinteraction atanaphase/cytokinesis.Phosphorylation of RhoGEFregulates this interaction. Thedata support a model in whichautoinhibition of RhoGEF isrelieved by GAP binding [2,3]

What Defines the Width of theFurrowing Zone?Live imaging of active RhoAshowed that a discrete, well-defined zone of RhoA activity iscritical for successful cytokinesis:when the zone is diffuse furrowingis initiated, but cleavage fails.Further, it is clear thatmicrotubules direct the formationof the zone in an active, dynamicway. How are narrow zones ofactive RhoA generated? Onepossibility is that kinesin-6transports GAP to microtubuleplus-ends adjacent to the cellcortex; Rho activators couldaccumulate in regions ofmicrotubule overlap, and diffuseaway elsewhere. This possibility isconsistent with the observation

that RhoA zones are diffuse whenmicrotubule organization isdisrupted, or when microtubulesare distant from the cortex.Bement et al. [1] further suggestthe interesting possibility that theGAP might inactivate RhoA-GTPthat is not bound to targetproteins, thus preventingbroadening of the RhoA zone.

In mammalian cells, where thecentral spindle pathway isdominant, centralspindlin clearlycontributes to signal delivery [9].However, furrows can be initiatedby either astral or central spindlemicrotubules, and in both casesRho is required [7]. Is transport ofactivators of RhoA on astralmicrotubules also driven bycentralspindlin? Or are othermechanisms utilized to locallyactivate RhoA? In mammaliancells, Aurora B can be delivered tothe cortex independently ofcentral spindle microtubules [11]and in fly embryos GFP-taggedPavarotti (kinesin-6) localizes toastral as well as interzonalmicrotubules [12]. Analysis of thedistribution and dynamics of thesignal delivery system in live cellswill help to resolve these issues.

Microtubules Are Both Positiveand Negative Regulators ofFurrowingThese new results are consistentwith microtubules delivering, andoverlap zones concentrating,positive inducers of cytokinesis.Other work, however, hassuggested that microtubulesdeliver signals that inducecortical relaxation. If thesesignals were maximal at thepoles, equatorial contractioncould occur [13]. The idea thatmicrotubules negatively regulatefurrowing is supported by theobservation that disruption ofmicrotubules by drugs or geneticmanipulations results in ectopicfurrows [5].

One possible way to reconcilethese different hypotheses is thatmicrotubules participate in bothpositive and negative regulationof the cortex. Perhaps dynamicmicrotubules exert a global,negative effect on contractility,and also serve as tracks todeliver activators of RhoA to thecortex. Regions of microtubule

overlap — or more generally,asymmetry in microtubuledistribution along the cortex [5]— could dynamically generatezones of active RhoA,overcoming negative regulationand stimulating contractility.Active RhoA could then locallystabilize microtubules [14].Although the idea is speculative,stable microtubules might in turnenhance delivery of RhoAactivators, generating a positivefeedback loop (Figure 1).Consistent with this possibility,regional differences inmicrotubule behavior have beendemonstrated in dividing culturedcells [15] and stable microtubulesconcentrate interzonalcomponents [16]. It will be ofinterest to determine if ectopicfurrows are associated withzones of RhoA activity.

The new capacity to directlyvisualize active RhoA duringcleavage has provided freshinsight into the mechanisms thatregulate cytokinesis.

References1. Bement, W.M., Benink, H.A., and von

Dassow, G. (2005). A microtubule-dependent zone of active RhoA duringcleavage plane specification. J. Cell Biol.170, 91–101.

2. Zhao, W., and Fang, G. (2005).MgcRacGAP controls the assembly ofthe contractile ring and the initiation ofcytokinesis. Proc. Natl. Acad. Sci. USA102, 13158-13163.

3. Yuce, O., Piekny, A., and Glotzer, M.(2005). An ECT2-centralspindlin complexregulates the localization and function ofRhoA. J. Cell Biol. 170, 571–582.

4. Rappaport, R. (1996). Cytokinesis inanimal cells, Volume 30, First edition(Cambridge, UK: Cambridge UniversityPress).

5. Glotzer, M. (2004). Cleavage furrowpositioning. J. Cell Biol. 164, 347–351.

6. Alsop, G.B., and Zhang, D. (2003).Microtubules are the only structuralconstituent of the spindle apparatusrequired for induction of cell cleavage.J. Cell Biol. 162, 383–390.

7. Bringman, H., and Hyman, A. (2005). Acytokinesis furrow is positioned by twoconsecutive signals. Nature 436,731–734.

8. Vagnarelli, P., and Earnshaw, W.C.(2004). Chromosomal passengers: thefour-dimensional regulation of mitoticevents. Chromosoma 113, 211–222.

9. Glotzer, M. (2005). The molecularrequirements for cytokinesis. Science307, 1735–1739.

10. Somers, G.W., and Saint, R. (2003).A RhoGEF and Rho family GTPase-activating protein complex links thecontractile ring to cortical microtubulesat the onset of cytokinesis. Dev. Cell 4,29–39.

11. Murata-Hori, M., and Wang, Y.-L. (2002).Both midzone and astral microtubule areinvolved in the delivery of cytokinesissignals: insights from the mobility ofaurora B. J. Cell Biol. 159, 45–53.

Dispatch R873

12. Minestrini, G., Harley, A.S., and Glover,D.M. (2003). Localization of Pavarotti-KLP in living Drosophila embryossuggests roles in reorganizing thecortical cytoskeleton during the mitoticcycle. Mol. Biol. Cell 14, 4028–4038.

13. Oegema, K., and Mitchison, T.J. (1997).Rappaport rules: cleavage furrowinduction in animal cells. Proc. Natl.Acad. Sci. USA 94, 4817–4820.

14. Palazzo, A., Cook, T.A., Alberts, A., S.,

and Gundersen, G.G. (2001). mDiamediates Rho-regulated formation andorientation of stable microtubules. Nat.Cell Biol. 3, 723–730.

15. Rusan, N.M., and Wadsworth, P. (2005).Centrosome fragments and microtubulesare released and transportedasymetrically away from division plane inanaphase. J. Cell Biol. 168, 21–28.

16. Canman, J.C., Cameron, L.A., Maddox,P.S., Straight, A., Tirnauer, J.S.,

Mitchison, T.J., Fang, G., Kapoor, T.M.,and Salmon, E.D. (2003). Determining theposition of the cell division plane. Nature424, 1074–1078.

Biology Department, 221 Morrill South,University of Massachusetts, Amherst,Massachusetts 01003-5810, USA.

DOI: 10.1016/j.cub.2005.10.021

Current Biology Vol 15 No 21R874

Arwen Long and Michael Platt

“George could not resist. Hesimply HAD to open it.”–Margret and H.A. Rey, CuriousGeorge Goes to the Hospital [1]

Self control is not just a problemfor curious little monkeys. Bothpeople and animals often preferimmediate over delayed rewards,even when patience would yield amuch bigger payoff [2–4]. Thisphenomenon, known as temporaldiscounting in economics andimpulsivity in psychology, impliesthat the value of the more distantreward is diminished, ordiscounted, by the timeintervening between the choiceand the reward. Temporaldiscounting helps explain why it isso difficult to get teenagers tosave for retirement or why it isimpossible to leave a dog at homewith enough food for a week andexpect it to ration itsconsumption.

Temporal discounting istypically studied by offering achoice between a smaller rewardnow and a larger reward sometime in the future. By changing thesize of rewards and their delays,the rate at which future rewardsare discounted can be calculated.Such estimates suggest that thevalue of future rewards declinesrapidly [2,5], and this decline canbe described mathematically by a

hyperbolic function [2] or the sumof two exponential functions [5,6].

At first blush, temporaldiscounting may seem irrational— why should a monkey preferone banana now when it couldhave three in a week? Discountingmakes more sense in light of theinherent uncertainty of futurepayoffs. Taking one banana todayis guaranteed, but a monkeymight not survive long enough toharvest a bumper crop of futurebananas.

While patience may be a virtue,“perseverance furthers” [7]. Ourcultural valuation of hard workhighlights the fact that manypeople prefer the easiest optionseven when they are lessrewarding, just as they valueimmediate over delayed rewards.As they reported recently inCurrent Biology, Stevens andcolleagues [8] have now shownthat monkeys also discountrewards by effort. In theirexperiments, two species of smallSouth American monkeys,common marmosets and cotton-top tamarins, were offered achoice between a small rewardand a large reward (Figure 1A).When the two rewards wereequally close, both tamarins andmarmosets chose the largerreward. As the distance to thelarger reward was increased, thetamarins continued to prefer it,but the marmosets abruptlyswitched their preferences and

chose the smaller, but closer,reward.

Stevens and colleagues [8]suggest that both species basetheir choices on the relative,rather than absolute, magnitudesof rewards. When given a choicebetween two close or six distantbanana-flavored treats, bothspecies chose the larger rewardwith the same frequency as whengiven a choice of one close orthree distant treats. Like otheranimals [3], tamarins andmarmosets seem to choosebased on the ratio of rewardsrather than their absolute values.The authors ruled out thepossibility that simple perceptualdifferences, such as an inability todiscriminate the larger rewardswhen they were far away, mightaccount for spatial discounting bymarmosets [8]. Indeed, as bothspecies have similarly sizedbrains, body sizes, and socialstructure, it seems unlikely thatany of these factors could beresponsible for the differences inspatial discounting reported.

These observations suggestthat marmosets discount rewardsas a function of distance or effort(Figure 1B), all the more surprisingin light of a separate set ofexperiments by Stevens andHauser [9] which showed thatmarmosets are more patient thantamarins when choosing betweenimmediate small rewards anddelayed larger rewards. In thoseexperiments, marmosetsoverwhelmingly preferred thelarger rewards — even when theywere delayed for longer than ittook to travel to the distant butlarger rewards in the new study(Figure 1C) [8]. These datasuggest that marmosetsdifferentially discount rewards asa function of time and effort.

A central tenet of economics isthat choosers show ordered,

Decision Making: The Virtue ofPatience in Primates

Marmoset monkeys devalue rewards requiring travel to acquire, buttamarin monkeys do not, despite the greater patience of marmosetswhen rewards are delayed in time. Such preference reversals, notpredicted by standard economic theory, may reflect behavioralmechanisms adaptively specialized for different spatial and temporalpatterns of foraging.