Technical Comments

Response to Comment on "Ivory-billed Woodpecker (Campephilus principalis) Persists in Continental North America"

Science  17 Mar 2006:
Vol. 311, Issue 5767, pp. 1555
DOI: 10.1126/science.1123581

This article has a correction. Please see:


Claims that the bird in the Luneau video is a normal pileated woodpecker are based on misrepresentations of a pileated's underwing pattern, interpretation of video artifacts as plumage pattern, and inaccurate models of takeoff and flight behavior. These claims are contradicted by experimental data and fail to explain evidence in the Luneau video of white dorsal plumage, distinctive flight behavior, and a perched woodpecker with white upper parts.

Since obtaining the Luneau video (1), we have treated the southern pileated woodpecker (Dryocopus pileatus pileatus) as the null hypothesis for the bird's identity. We analyzed 56 videos of pileated woodpeckers launching and flying, and we tested the hypothesis experimentally with models. Numerous features of the bird differ substantially from those of the pileated woodpecker and match the ivory-billed woodpecker (Campephilus principalis), forcing us to reject the null hypothesis. We address the main points of Sibley et al. (2) here and present a comprehensive analysis of the Luneau video elsewhere (3).

The case presented by Sibley et al. (2) contains the following problems: (i) It relies on erroneous representation of the underwing of pileated woodpeckers as mostly white with a narrow black rim [figures 1C and S2 in (2) and recent field guides by the lead author (4)]. Pileated wings have more black than white on the ventral surface [see supporting online material (SOM) text]. (ii) Proposed wing and tail movements during the woodpecker's launch do not match the video images. (iii) Sketched renditions of video fields incorporate video artifacts as plumage pattern and are unsupported by experimental or field documentation. (iv) It dismisses experimental demonstration that the observed underwing pattern matches a flapping model of an ivory-billed, not a pileated. (v) It invokes an extraordinary model of flapping flight whereby a bird in caudal view conceals dorsal wing surfaces on both downstroke and upstroke. (vi) Without supporting evidence, it dismisses additional features of the Luneau video that are consistent with ivory-billed and inconsistent with pileated.

White wing patch at launch. Contrary to (2), the pattern of white in field 33.3 matches an ivory-billed woodpecker wing viewed laterally from slightly below the horizontal (Fig. 1). Its size is fully consistent with an ivory-billed woodpecker (3) (SOM text).

Fig. 1.

Luneau video field 33.3 (A) shows the first major appearance of the woodpecker's right wing to the left of the tupelo trunk. The white triangle has black above and a small black spot between the lower edge and the trunk. We interpret this as a lateral view of the opening wing of an ivory-billed woodpecker as the bird begins to turn away from the approaching observers, matched here by a montage (B) of a specimen's wing (C) superimposed behind a tupelo trunk. Sibley et al. (2) propose, instead, that the pattern is a vertically extended underwing of a pileated woodpecker (D), but comparison with a pileated woodpecker wing specimen at such an angle (E) reveals flaws in their diagram. A pileated woodpecker wing would show a broad black border entirely encircling the white and comprising 60 to 70% of the wing area. If the wing were tilting away from or toward the viewer, thus showing less black, the white underwing also would be extremely narrow (F). Moreover, the position of the tail and body proposed by Sibley et al. (2) in this field are incompatible with the sequence of movement observed in adjacent frames.

The pattern in fields 16.7, 33.3, and 50 does not match a pileated woodpecker wing (Fig. 1). Sketches in (2) show a black rim around the edge of the white, but no such black appears in the actual video fields (grayish background areas are not associated with the wings). If the white were a pileated underwing, then unambiguous black should be present distally on all remiges viewed ventrally from any angle (Fig. 1) (3).

The statement that when woodpeckers launch “the initial wing extension happens very quickly and, given the angle at which the bird is viewed, the underwing should be visible” (2) is inconsistent with video evidence. Woodpecker launches are variable (3) but typically begin with the wings opening slightly as the body begins to pivot (SOM text). Sketches (2) of a bird suddenly airborne with wings outstretched overhead, body already lateral to the camera, yet tail and body still nearly vertical, are atypical. Even the videos that Sibley et al. cite (5) demonstrate our point.

Wings largely white, wingtips black. The crucial claim that during flight “the white on the wings can be accounted for by the ventral surface” (2) is incorrect, and the “wing-twisting” hypothesis contradicts all models and photographic analyses of flapping flight in birds (68). Again, when viewed in full, even the images and videos (5) Sibley et al. cite reveal dorsal wing surfaces visible caudally.

Ventral wing surfaces of a flying woodpecker viewed caudally are sometimes visible during mid-downstroke, but on the upstroke and immediately before the downstroke the dorsal surface prevails (68). Referring to the distinctive Luneau video, Sibley et al. (2) state that “lack of an obvious black trailing edge in most video frames does not rule out an identification of pileated woodpecker,” but no video evidence exists to support this claim. To the contrary, the black portion of the ventral wing surface (>50% of the wing) (SOM) is visible in virtually every frame of a flying pileated woodpecker (Fig. 2) (3). The claim by Sibley et al. (2) that in certain fields the black wingtips of the woodpecker show a curved shape suggestive of pileated fails to acknowledge that movement blur is most pronounced at the wingtips and that in other fields of the Luneau video the shape is more suggestive of ivory-billed.

Fig. 2.

Effects of video artifacts on wing patterns of pileated woodpecker and ivory-billed woodpecker. To reenact the Luneau video (1), life-sized models with flappable wings were painted to resemble pileated woodpeckers (A) and ivory-billed woodpeckers (B). Deinterlaced video fields of these models shot with the same camera, distance, and light conditions as the Luneau video reveal the prominent black edge on the pileated woodpecker model (C). The ivory-billed woodpecker model shows a white underwing with black wingtips, the central black wing line being lost because of white dominance and poor video resolution (D). Our interpretive sketches of these fields differentiate the clear black trailing edge of the pileated woodpecker (E) from indistinct dark borders of the ivory-billed woodpecker model (F), which are video artifacts. Sibley et al. (2) erroneously interpreted such artifacts in the Luneau video as a thin, black trailing edge. Deinterlaced, poor-resolution video fields of pileated woodpeckers in flight (G, I, and K) show wing patterns resembling the pileated woodpecker in the reenactment. Deinterlaced video fields from the Luneau video (H, J, and L) show wing patterns resembling the ivory-billed woodpecker in the reenactment.

In the Luneau video, white dominates the wings throughout, including many unambiguously dorsal views (SOM). The left wing is white to its base as it emerges from behind the tree [field 250 and subsequent fields (3)] and in additional fields that show dorsal views during the latter stages of upstrokes (e.g., fields 700, 816.7, and 950). The right wing shows dorsal white as the bird veers left. Contrary to Sibley et al. (2), white is evident in many fields in which the wings are closing or folded during the upstroke (3), whereas similar positions in pileated woodpecker videos show all-dark wings. Extensive white along the rear dorsal surface, conspicuous black wingtips on every wingbeat, and absence of black along the rear edge of the wing eliminate pileated and represent the diagnostic wing pattern of an ivory-billed woodpecker.

We conducted experimental reenactments using ivory-billed and pileated models designed to demonstrate the full wing extension at the beginning of the downstroke and videoed these at shutter speeds slow enough to simulate conditions of the Luneau video (Fig. 2) (3). The ivory-billed model yielded images strikingly similar to those in the Luneau video, but comparable images of a pileated model revealed the expected black trailing edge and were incompatible with the Luneau video.

White on dorsum. We regard the presence of white on the back of the Luneau woodpecker as obvious and indisputable (SOM text). The statement that “little of its back is visible” (2) is irrelevant, because some white on an ivory-billed's back would be visible at virtually any caudal angle. The amount and placement of white is inconsistent with its being on the head and neck, which pointed directly away from view throughout. Reflection off a pileated's sooty-black back on a cloudy day cannot produce the persistent whitish areas visible at several angles. Finally, although resolution is too poor to identify a double stripe unambiguously, the suggestion of such exists in field 866.7.

Rapid and direct flight. The Luneau woodpecker flies with a wingbeat frequency of 8.6 Hz without undulation for more than 4 s. The 1935 audio recording of a pair of ivory-billed woodpeckers at a nest (SOM text) captured one bird flying away with noisy wingbeats [as described in (9)] having a frequency of 8.4 Hz (3). The close match between the Luneau woodpecker and the 1935 recording is especially important because both are faster than any wingbeat frequency ever documented for pileated woodpecker. The sustained duration of this direct flight pattern by the Luneau woodpecker is extraordinary, because pileated woodpeckers typically shift to slower, deeper wingbeats moments after launching from a perch, even when the initial few beats are rapid (3).

Perched woodpecker. Contrary to the interpretation of Sibley et al. (2), the black and white object apparent in the Luneau video 26 s before the bird flies is consistent in size (35 to 45 cm), shape (vertically elongate, leaning away from the trunk), and pattern (black with white central patch) with a perched ivory-billed woodpecker [video close-up in (3)]. The object remains fixed on the trunk as the camera's viewing angle shifts relative positions of objects at different distances from the camera (fig. S2). Unlike the objects identified in figure S1 of Sibley et al. (2), this object is too large, too well-defined over a 6-s period, and its midportion too white to be a video artifact or leaf cluster. Moreover, the object was gone when the Luneau canoe came around the bend in the bayou and the woodpecker launched into flight nearby (fig. S2). It was never present on subsequent examinations of the site. Identity of this object is not crucial to identification of the flying bird, but we cannot explain its size, pattern, and disappearance from view except as an ivory-billed woodpecker that flew 3 m to hide behind a tupelo, then fled moments later as the canoe approached. No evaluation of the Luneau video can be considered exhaustive without a credible interpretation of this feature.

Interpretations of frames in the Luneau video by Sibley et al. (2) portray wing patterns, video artifacts, and wing twisting that accord neither with experimental and comparative findings nor with models of bird flight. Video (1), sightings (10, 11), and suggestive acoustic evidence (12) establish that at least one ivory-billed woodpecker persisted in eastern Arkansas during 2004 and 2005, compelling vigorous search and conservation efforts in the Big Woods and elsewhere across the southeastern United States.

Supporting Online Material

SOM Text

Figs. S1 to S3

References and Notes

References and Notes

View Abstract


Navigate This Article