The Bowerbirds: Ptilonorhynchidae.— Clifford B. Frith and Dawn W. Frith. 2004. Oxford University Press, Oxford. xxiii + 508 pp. ISBN 0198548443. Paper, $164.50.—This 10th Oxford Bird Families of the World volume treats the 20 species of bowerbirds endemic to Australia and New Guinea. The unique and fascinating behavior of building and decorating bowers has brought much attention to this family, for which there has been no serious book-length treatment since Gilliard's 1969 classic. The first of the two parts of this book addresses general topics, including an introduction to the bowerbirds, biogeography, foraging and ecological cycles, morphology, demography, relationships among species, breeding biology, and parental care. Three additional chapters discuss issues specific to bowerbirds, such as bower site, bower structure, courtship display, and the evolution of bowers and the bowerbird mating system. Part two presents species accounts organized by the eight bowerbird genera. Between the two parts are pictures of bowerbird males at bowers and females at nests and a set of excellent illustrations of each species by Eustace Barnes.

Frith and Frith present a detailed review of the bowerbird literature, including many obscure references and personal communications. This is supplemented by numerous helpful tables, figures, and maps that provide information on morphometrics, bower dimensions, numbers of decorations, nest characteristics, egg dimensions, biogeography, and other aspects of bowerbird natural history. The authors discuss the phylogenetic placement of the bowerbirds in relation to other families, and the relationships of bowerbird genera and species, using morphological, behavioral, and limited molecular information. Classification at the species level is little changed from recent descriptions, with the exception that the two morphologically distinct Flame Bowerbird subspecies are now separated as the Masked Bowerbird (Sericulus aureus) and the Flame Bowerbird (S. ardens).

The authors' informal and qualitative analysis of species' relationships is biased toward morphological characters, including plumage, but they do not explicitly consider behaviorally based traits such as bowers and decorations. They fail to use modern phylogenetic techniques, and this may contribute to their inability to resolve problematic issues associated with species' relationships. The robust Kusmierski et al. (1997) mithochondrial-DNA bowerbird phylogeny is not shown or used in any systematic attempt to resolve relationships among bowerbird species. Significantly, that phylogeny suggests a major change in the relationships of the bowerbirds: the Streaked Bowerbird (Amblyornis subalaris) is a sister group to the clade including Prionodura, Archboldia, and the remaining Amblyornis, which suggests that those taxa form a single genus. And although they suggest that those taxa are tightly linked, Frith and Frith offer no clear evidence supporting that claim.

It is hard to read anything written about bowerbirds and not be captivated by their amazing natural history, but serious errors and omissions greatly detract from this volume. Figure 1.1 is meant to show the distribution of bowerbirds but omits the distribution of Satin and Regent bowerbirds along the Australian east coast. In Table 4.1, values given in the third row in each cell do not show male-female differences as stated. In numerous instances, work of other researchers is inaccurately represented. Frith and Frith (p. 123) mistakenly report that Borgia et al. (1987) found that blue items were common in the habitat, when actually we showed the opposite; and they report that only one later paper quantitatively assessed likely decorations in the habitat, when our paper had used the same methods. Too often, Frith and Frith cite papers without mentioning that they are about nonbowerbird species and have dubious relevance to bowerbirds; for example, it is unclear how Mulder and Magrath's (1994) finding that fairy wrens use the timing of prenuptual molt to assess male quality is useful in understanding bowerbird mate assessment, which occurs at bowers when all males have completed their molt.

The authors' views are commonly offered without support or with explanations that are less than compelling. They reject others' observations that males trim leaves over bower platforms to increase light falling on the bower platform (p. 103), even though (1) they have observed males trimming leaves, (2) they argue that light on platforms likely critically affects display, and (3) it is clear that removing leaves that block light just above the bower platform would change its illumination. They recognize that divergence in female preferences between Vogelkop (Amblyornis inornatus) populations leading to dramatic differences in bower shape and decoration “supports the speciation by sexual selection hypothesis” (p. 29) but, without explanation, conclude that those populations have not attained species or even subspecies status (Table 2.4). Frith and Frith uncritically report the claim that local similarities in bower decoration are attributable to cultural transmission, even though there has been no attempt to measure cultural transmission and likely alternative hypotheses were not effectively eliminated.

The discussion of the evolution of bowers and bower decorations (Chapter 7) deserves special attention, because these are the signature traits of the bowerbirds. Frith and Frith support our (Borgia et al. 1985) marker hypothesis, which suggests that bowers and decorations serve as indicators of male genetic quality. But that hypothesis requires a pre-existing array of bowers with differences related to genetic male quality, so another hypothesis is needed to explain bower origins. Frith and Frith propose as “plausible” the hypothesis that bowers evolved to hide the female during courtship from interfering males. That hypothesis receives little behavioral support from species of the two major clades of bower-building species. In the maypole-building clade, maypole or courts without bowers almost fully expose the female and, in the two independently evolved hut-building species, the courting male hides in the hut while the female stands outside. Avenue bowers expose females from the front, rear, and above, the directions of likely approach by interfering males; and females of several species often do not move completely into the bower avenue, apparently not concerned with hiding. Frith and Frith suggest that over-hanging bushes hide the female from above, but that is rarely complete, and therefore does not prevent intruding males from determining whether a female is in the bower. Also, if bushes typically completely hide the female, that would eliminate the need for bowers rather than favor their evolution.

Frith and Frith dislike my hypothesis that bowers originated to increase female comfort when visiting display courts in response to the threat of forced copulation by the courting male (I call this the “female protection from forced copulation” (FPFC) hypothesis, but Frith and Frith incorrectly label it the “rape hypothesis”). They question how males can benefit from building a structure that prevents their own forced copulation, ignoring my suggestions that bower-building males benefit from (1) increased female visitation by less-threatened females and (2), for the best males, less-threatened females who are able better to attend to and choose their high-quality displays. I suggested that males initially courted from behind trees, thus enhancing the female's opportunity to escape male attempts at forced copulation. Males then evolved to embellish those obstacles with sticks (to form true bowers), allowing improved functional design of bowers over naturally available barriers. Supporting the FPFC hypothesis is (1) strong evidence that females are under threat of forced copulation, (2) evidence that the hypothesis is functionally consistent with existing bower types (in all bower-building species, males must run around the bower wall or maypole to mount the female, thus giving her time to escape unwanted copulations), and (3) the requirement to deal with the threat of forced copulation in court-clearing bowerbirds that do not build bowers. In the non-bower-building Archbold's Bowerbird (Archboldia papuensis), the unique male groveling courtship behavior is consistent with reducing threat of forced copulation but is not predicted by the hiding-from-the-female hypothesis that Frith and Frith prefer. In Toothbilled Bowerbirds (Scenopoeetes dentirostris), there also is no bower; and after very short displays, males typically capture visiting females by their nape in violent copulations that are similar to forced copulations.

The authors (p. 219) present two quotations from me suggesting contradictory claims about the role of the bower in preventing forced copulations, but the second was taken from the discussion of a different hypothesis suggesting that the bower protects females from aggressive physical attacks by males. They first deny (p. 219), then accept (p. 220) the occurrence of forced copulations, and then argue that females are not susceptible to forced copulation because they “turn onto their back to defend themselves” (p. 220). But our extensive video records show that females do not fight males and that birds on their back are typically intruding males attacked by the bower owner. Frith and Frith claim that forced copulations by marauding male Satin Bowerbirds (Ptilonorhynchus violaceus) are unimportant because they are carried out principally by sexually immature males, when in fact such copulations involve five-year-old and older juvenile-plumage males, which Marshall (1954) showed make viable sperm. Also, they never consider why those males carry out this risky behavior that may end in prolonged and presumably costly fights with bower owners if there were no benefit. Frith and Frith point to the complex system of male and female signaling we found in Satin Bowerbirds as evidence against a threat of forced copulation, missing the point that top males can gain more by courting rather than forcing copulation, and that bower-building and courtship-signaling are similar in allowing males to reduce the threat to females to improve their chances for mating. Curiously, Frith and Frith seem to lose track of the hypothesis they are attempting to refute, confusing the role of the bower in protecting the female from forced copulation by the courting males with other, different hypotheses (i.e. protection from forced copulation by marauding males and protection from male aggression). In an attempt to refute the role of forced copulation in shaping bowerbird behavior, they ask why it is not important in other birds, not considering the obvious fact that female bowerbirds are unique in surveying decorated ground courts where they are more susceptible to forced copulation. Apart from discussing forced copulation, Frith and Frith fail to consider the other evidence offered to support the protection hypothesis.

The authors also reject the “threat reduction” hypothesis; they suggest that it was proposed as a general hypothesis for bower evolution, even though I developed it specifically to explain the unique bowers and high-intensity displays of some Spotted Bowerbirds (Chlamydera maculata). I suggested that males build a unique thin, see-through bower wall to filter out threatening aspects of their otherwise attractive high-intensity displays. Without having seen a bower in the population in question, Frith and Frith state, “Certainly a male spotted could burst through his wall to access the female” (p. 221). We have monitored on videotape more than 2,300 courtships in this population and have never seen males attempt to burst through the wall, and if they could, the delay associated with passing through the wall would likely allow females to escape. Frith and Frith do not mention our experiments in which we destroyed one bower wall and males courted predominantly from behind the remaining wall and lowered their display intensity when on the side without the wall, which supported the hypothesis that males use the wall to filter intense display elements. They propose that females may favor male aggressiveness because it is a signal that they are better able to defend bowers and then pass on this trait to offspring. But that does not explain the unique behavior of displaying through specially modified bower walls and the adjustments that males make in our wall-destruction experiments.

The bowerbirds represent one of the high points of avian evolution and as such they deserve a book that fully captures the wonder of their fantastic natural history. In some ways, this book reflects many of the exciting aspects of bowerbird evolution. With its many references, maps, and tables, this is a useful tool for gaining access to the bowerbird literature. But what is most exciting about these birds is to understand how they came to have and use their unique adaptations—bowers and decorated courts. This book's quirky review of current thinking represents a missed opportunity to tell this exciting story, or at least what we know of it so far. This and the too-often inaccurate presentation of factual information suggest caution in using this book as a definitive source, particularly for issues related to bower evolution.