Ways to Raise Tadpoles Environments

NEWS & VIEWS NATURE|Vol 464|15 April 2010

VISION dark images against a background of bright sky, such as the silhouette of TEMPLE . Fisheye views an aerial predator. S The ventro-temporal retina Archerfish eyes must cope with in other species — the pigments presents peak sensitivities at the light spectrum of two very and spectral tuning vary between 453 nm, 535 nm and 565 nm. This different media. These fish live different parts of the retina (in this combination is possibly used for among mangroves and in rivers, case, subdivided into dorsal, colour vision along the visual axis and are renowned for their ability ventro-nasal and ventro-temporal appropriate for sighting prey against to bring down an insect target from regions). They interpret these a background of foliage. From video overhanging foliage by spitting a variations in terms of the tasks the recordings, Temple et al. conclude stream of water at their prey, as eye has to perform when operating that this part of the retina aligns with pictured here. Shelby Temple and along three visual axes; that is, spitting angles, and they estimate colleagues have investigated the three directions in which the eye that the visual resolution allows visual pigments and spatial resolving might look. an archerfish to tell the difference power of archerfish eyes, and A simplified description of their between two objects 2 mm apart at present the results in the context of conclusions is that the dorsal retina a range of 550 mm. the requirements for vision at the has maximum spectral sensitivity Archerfish can be trained to spit at water–air interface (S. Temple et al. at 454 and 570 nanometres, a beneath them. The values for the coloured targets. That, say Temple Proc. R. Soc. Lond. B doi:10.1098/ combination that Temple et al. ventro-nasal retina (visual axis et al., makes them excellent subjects rspb.2010.0345; 2010). consider is well suited for up and behind) are 502 nm and for investigating further aspects of the Using microspectrophotometry, discriminating between shades of 620 nm. The 502-nm peak is tuned, function of intra-retinal differences. the authors find that — as is known brown, and for identifying objects the authors suggest, for detecting Tim Lincoln

20 of the mutations was comparable for all three part in the pathophysiology of the tumour and initial clues to the most useful candidates for tumours, that 26 showed increased prevalence metastasis, and provide clues about the biologi- functional assessment. ■ in the xenograft and/or the metastatic tissue, cal roles of these genomic regions. They could Joe Gray is in the Lawrence Berkeley National and that the prevalence of 2 was significantly also provide information about clonal diver- Laboratory, Life Sciences Division, Berkeley, decreased relative to the primary tumour. This sity in metastatic lesions, which may help to California 94720, USA. suggests that at least three cell clones from the identify subpopulations of cellular molecules, e-mail: [email protected] primary tumour carried over into the meta- and thus influence cancer therapy. static and xenograft tumours: one carried Ding and colleagues’ data therefore hint that 1. Mardis, E. R. & Wilson, R. K. Hum. Mol. Genet. 18, R163–R168 (2009). mutations that decreased in prevalence, one future sequencing of some metastatic-cancer 2. Metzker, M. L. Nature Rev. Genet. 11, 31–46 (2010). had mutations that increased in prevalence, genomes should be considerably deeper than 3. Morozova, o., Hirst, M. & Marra, M. A. Annu. Rev. Genomics and one carried mutations whose prevalence contemplated at present to allow statistically Hum. Genet. 10, 135–151 (2009). did not change compared with the primary robust estimates of mutation prevalences to be 4. Ding, L. et al. Nature 464, 999–1005 (2010). 5. nguyen, D. X., bos, P. D. & Massagué, J. Nature Rev. Cancer tumour. This metastasis therefore does not obtained. More over, functional assessments of 9, 274–284 (2009). seem to have formed from a single cell, but the affected genomic regions will be needed to 6. gupta, g. P. & Massagué, J. Cell 127, 679–695 (2006). rather from a cell population that, in this case, determine which genes in the selected clones 7. Denardo, D. g., Johansson, M. & Coussens, L. M. Cancer contained at least these three clones. are drivers and which mere passengers. Con- Metastasis Rev. 27, 11–18 (2008). 8. Langley, R. R. & Fidler, I. J. Endocr. Rev. 28, 297–321 (2007). Another remarkable result was that 16 of the cordant selection of mutations in the meta- 20 mutations present at increased prevalence in static and xenograft tumours could provide See also News Feature, page 972. the metastatic tumour were also present with higher prevalence in the xenograft. This pat- tern of concordant selection of one or more clones carrying common mutations during BEHAVIOURAL ECOLOGY progression to metastasis and establishment of a xenograft suggests that similar evolutionary pressures are at work on these cells in both Ways to raise tadpoles environments. This may provide some indica- Hanna Kokko and Michael Jennions tion of the aspects of the metastatic process that are influenced by the aberrations carried To reduce parental care, just add water — that’s the conclusion of an in the selected clone, because some processes intriguing investigation into the extent of the motherly and fatherly needed for metastasis, such as invasion and barrier penetration, may not be important devotion that different species of frog extend to their offspring. selective forces in the xenograft environment. These aspects of metastasis therefore may not Nature documentaries frequently invite their ranging from protecting them against preda- be influenced by aberrations that are selected viewers to contemplate that only a tiny minor- tors and environmental stresses to giving them in both xenograft and metastasis. ity of the perfectly formed larvae floating in shelter or food. Of course, Ding et al.4 assessed the evolu- the sea, or crawling on land, can ever hope to Why does this diversity of solutions exist? tion of only one tumour. But if their results can make it to the adult stage. Parents of many spe- Theoreticians state that parents may reduce be reproduced in larger studies, comparative cies seem surprisingly unconcerned, leaving care to increase the number of young produced investigations of primary-tumour/xenograft/ their progeny to fend for themselves. In some and/or to improve the parents’ own survival. metastasis triplets might facilitate identification species, however, parents provide their off- Both trade-offs reduce the lifetime fecundity of genomic aberrations that play a significant spring with costly, time-consuming services, of parents, so parents are likely to provide care

only if it substantially improves the survival of frogs is rare, so factors other than the need W

1 o

the offspring they do produce . A study of frogs to feed offspring are probably required to bR by Brown et al.2, just published in The Ameri- explain the patterns of parental care in frogs on L. on

can Naturalist, shows that a species’ breeding more fully. For example, breeding in a small AS J habitat can influence offspring survival in a body of water could decrease the likelihood of surprisingly deterministic way: tadpoles have brood parasitism (in cases in which more than a chance of surviving on their own only if they one pair are using a breeding pool), because live in relatively large pools of water. later-hatching tadpoles are often cannibalized It has proved difficult to identify specific by those that hatched earlier4. In addition, the ecological factors that affect whether parental use of a small breeding ground increases the care will evolve, yet the answer provided by certainty of genetic parentage, making it more Brown and colleagues2 is simple and elegant. likely that parental care will evolve. It also To start with, they built a phylogeny depicting influences which sex provides care — males, the evolutionary relationships among 404 frog females or both parents5 — but evolutionary species that are distributed across the order transitions do not occur as easily in all direc- Anura and that have been investigated for tions6. It might not be a coincidence that the whether and how they provide parental care. poison frog species that engages in biparental Frog species differ enormously in whether care is also the one that is genetically more they care for their young and in the type of monogamous. care given: parents may protect eggs by lay- Even so, caution should be exercised in using ing them in terrestrial burrows, or brood the the argument that monogamy is conducive to young in pockets of tissue on their back or in biparental care and therefore concluding that the mother’s stomach3. Although Brown et al. circumstances that make offspring needy also did not consider all frog species (of which favour monogamy. The co-evolution of traits there are more than 5,300), their phylogeny tends to occur with a delay. Thus, the first uncovered compelling ecological generalities. monogamous parents of a hypothetical spe- Species that deposit eggs and tadpoles into cies might not reap any benefits from their Figure 1 | Tadpole transport. phyto telmata (small pools of water found in After tadpoles have behaviour as they will not yet have responded emerged from eggs, males of both the variable hollows in plants) are more likely to provide poison frog (Ranitomeya variabilis; top) and to the novel conditions of reliable parentage by parental care than those breeding in terrestrial the mimic poison frog (Ranitomeya imitator; switching to a more intensive form of care. It waters such as streams, ponds and rivers. The bottom) carry these offspring on their back from may be, however, that home-range size or other estimated rate of evolutionary transitions from the place of birth in a phytotelma (a water-filled ecological factors make monogamy more likely providing no care to providing care was nine hollow in a plant) to another phytotelma. in some species and that this later selects for times higher in species that breed in these tiny Brown et al.2 find that these frogs otherwise care by both parents. pools than in those that use terrestrial waters. differ considerably in the amount of care Regardless of the details of the final story, Why does a limited water source turn frogs provided by the parents. The smaller there is little doubt that frogs, with the diverse into devoted parents? One possible answer is phytotelmata favoured by R. imitator are ways in which they care for their offspring associated with a greater degree of parental that the food supply that a small body of water attention being paid to the tadpoles. and an increasingly well-resolved phylogeny, offers is so meagre that parents using these as are becoming an important group in helping breeding grounds have been strongly selected to explain why there is so much variation in to improve the survival of their offspring. To which is promptly eaten by the tadpoles. parental care among animals7. Unfortunately test this idea, Brown et al. investigated one Tadpoles of both species are thus cared though, this diversity continues to be lost. of the evolutionary contrasts in their data set for, but tadpoles of the variable poison frog For instance, the details of the ecology of the in detail, reporting fascinating differences develop without being fed and receive care only two species of gastric-brooding frog will between two species of poison frog that live in from only one sex (males). Is it a coincidence never be known, as both became extinct in the same habitat in Peru. that variable poison frogs use larger breeding the mid-1980s. ■ Individuals of the variable poison frog spe- pools than mimic poison frogs, or did feeding Hanna Kokko is in the Department of Biological cies (Ranitomeya variabilis) have large home evolve in mimic poison frogs specifically to and Environmental Science, PO Box 65, ranges, and both sexes frequently switch mat- combat a low food supply in tiny pools? University of Helsinki, Helsinki 00014, Finland. ing partners. Eggs are laid above a suitable To shed light on this question, Brown et al.2 Michael Jennions is in the Ecology, Evolution and phyto telma, and either the tadpoles fall into carried out a translocation experiment. They Genetics Division, Research School of Biology, the water or, more commonly, the male parent show that tadpoles of both species grow and Australian National University, Canberra, returns and helps the tadpoles to rupture the survive poorly in small pools when they are ACT 0200, Australia, where H.K. is currently egg membrane. He then transports the tadpoles denied parental attention, whereas such prob- a visiting fellow. on his back (Fig. 1) to another hollow. In this lems do not arise in larger pools. The fact that e-mails: [email protected]; species, males place tadpoles in water volumes tadpoles of both species suffered similar fates [email protected] averaging about half a cup (112 milli litres). In is crucial to the interpretation of the experi- the closely related, similarly sized mimic poi- ment. It allowed Brown et al. to circumvent the 1. Klug, H. & bonsall, M. b. Evolution 64, 823–835 (2010). son frog (Ranitomeya imitator), mating pairs chicken-and-egg problem that species in which 2. brown, J. L., Morales, V. & Summers, K. Am. Nat. 175, are often monogamous. Their home ranges are care is routinely given might have evolved 436–446 (2010). small, and males carry the tadpoles to a tiny tadpoles that obligately rely on parental care, 3. Wells, K. D. The Ecology and Behavior of Amphibians (Univ. Chicago Press, 2007). phytotelma which is, on average, the volume which could have led to an overestimation of 4. brown, J. L., Morales, V. & Summers, K. Biol. Lett. 5, 148–151 of half a shot glass (24 ml). The male parent the value of parental care in small pools. (2009). thereafter returns periodically to monitor The phylogeny constructed by Brown et al. 5. Kokko, H. & Jennions, M. D. J. Evol. Biol. 21, 919–948 (2008). whether it is time to feed the tadpoles. If he also reveals that the production of trophic eggs 6. Kokko, H. Ecol. Lett. 2, 247–255 (1999). calls incessantly, the female makes her way to is generally associated with breeding in phyto- 7. Summers, K., McKeon, C. S. & Heying, H. Proc. R. Soc. the pool and lays a specialized ‘trophic’ egg, telmata. However, trophic egg production by Lond. B 273, 687–692 (2006).