Torrent Ducks 33

Plumages and wing spurs of Torrent Ducks Merganetta

armata
MILTON W. WELLER 1

Introduction Because of the difficulties of studying

The Torrent Duck Merganetta armata of this species in the wild or in keeping
the Andean highlands is one of the most captive birds, an attempt has been made
colorful and intriguing of all species of here to utilize the specimens collected by
waterfowl. It has adapted to the demand­ scientists and professional collectors
ing environment of cascading mountain throughout the Andes. Pooling data from
torrents, and only rarely is seen in high­ these museum specimens permits the
land lakes. Taxonomically, the species has biologist to visualize and even quantify
been a puzzle, and Johnsgard (1966) has many of the major patterns before
recently reviewed its biology and tax­ attempting to solve problems in the field.
onomy. Both he and Delacour (1956) Such museum study is still impossible
recognize only one species but differ with rarer species and at one time would
in the number of races accepted. Sub­ have been difficult for Torrent Ducks.
specific names here refer to Delacour’s However, there is now an excellent col­
terms, which were based on Conover’s lection in the American Museum of
(1943) classification and which are re­ Natural History and an especially fine
corded on most museum specimens. series of young birds in the Conover
My own experience with the species in Collection at the Field Museum of
life is restricted to observations of several Natural History in Chicago. The sequence
pairs in Chile (Curicó Province) and ob­ of plumages was studied mostly from the
servation of its habitat in northern Pata­ latter collection but nearly 200 skins were
gonia (Provinces of Río Negro and Neu­ examined in museums in the eastern
quén) and subtropical Argentina (Pro­ United States.
vince of Jujuy) during 1964 and 1965. Another unusual feature of the Torrent
Subsequently, this species was chosen as Duck is the presence of a metacarpal spur
part of a study of the plumages of selec­ on each wing. True spurs are well-defined
ted Neotropical anatids because of its pointed projections of a bony core covered
dramatic sexual dimorphism despite its with horny material (as opposed to the
isolation from other ducks (see Sibley bony knobs on the wings of geese) and are
1957 for a discussion of factors possibly not common among Anseriformes. They
influencing dimorphism). occur in the three species of Screamers

1 Journal Paper No. J-5885 of the Iowa Agriculture and Home Economics Experiment
Station, Ames, Iowa. Project No. 1504.
34 Wildfowl

(Anhimidae), and in the well-named garleppi and ‘coffee-colored’ eyes in

Spur-winged Goose Plectropterus gam­ other races. Both sexes have wing spurs
bensis (Rand 1954). It is interesting that but the colors of these in life have not
Johnsgard (1966) has suggested placing been recorded.
the Torrent Duck in the Tribe Cairinini, The downy young are generally dark
which also contains the Spur-winged gray above with a narrow white dorsal
Goose. It is known that Screamers and stripe and two white bars on either side.
Spur-winged Geese fight with their wings They are white ventrally and on the face
but no such fighting has been reported and have a gray eyeline and an indistinct
among Torrent Ducks. dark patch in the auricular area.
During examination of specimens, it The juvenile plumage is female-like dor­
became obvious that spur length differs sally but is pure white below in both
with sex and age of the birds, as noted sexes. Some individuals, especially of
also by Conover (1943), Rand (1954) and Merganetta a. armata, have an orange-
Delacour (1956). Spurs were measured as cinnamon tinge to feathers in the neck
a potential age criterion to aid in under­ or upper-chest region. Apparently this
standing the chronology and sequence of occurs in both sexes although a con­
plumages. spicuous shortage of immature females
in collections makes me question the data
Description of plumages on some labels. The feathers of the side,
The colorful plumages of the various flank and rump are broadly barred.
forms of Torrent Ducks have been des­
cribed by Phillips (1926), Conover (1943) Sequence of plumages
and Delacour (1956), and a general des­ Because the sequence and number of
cription of the species will suffice here. plumages do not seem to differ in the
Adult males are characterized by a bold various races, comments here are for the
black-and-white head pattern and a species as a whole. The natal plumage is
brownish-gray body, wings and tail. The replaced in a pattern similar to that of
crown, nape and the area at the base of other ducks. The tail is renewed very
bill are black, and the face is white with early and some natal down clings to the
an eyeline extending down the neck to the tail-feathers until the tail is fully juvenile
upper breast and connected to the nape and much of the body is feathered. Natal
by another black stripe. In the southern down also clings to back feathers. The
Merganetta a. armata, the eyeline con­ advent of juvenile feathers on the ventral
tinues in a circular fashion forward and surface and side precedes those on the
downward and connects to a black chin, head and back. The entire body is essen­
throat and ventral neck. The back and tially fully feathered before the primaries
scapular feathers are elongated and develop noticeably. Although body
pointed, and various races have brown, plumage precedes wing development in
tan, gray or white edges to these feathers. most species, flight usually is possible
The chest, belly and side are very dark before or just as the back is fully
and rufous in the southern race, but feathered (Weller 1957). Presumably their
others have pale-gray breasts streaked diving ability and the habitat provides
with black. adequate safety and complete protective
Adult females also are grayish dorsally body feathering is more valuable than are
and on the head and neck, but they have flight feathers.
finely vermiculated feathers in the eyeline Replacement of juvenile by adult
area that extend down the side of the plumage begins as soon as the primaries
neck and along the side. The lower face, are fully grown, as evidenced by five
chin, throat and entire ventral surface are flightless juveniles with no tail molt
rich cinnamon. (Field Museum 10400, 14295, 16068,
An iridescent green speculum is 13762, 14491) and seven flying immatures,
present on the secondaries of both sexes which have some notched juvenile and
at all ages. It is bordered by a narrow some new adult tail-feathers (F.M. 13761,
white line anteriorly (on the greater 14962, 88246, 16097, 12263, 17432, 14298).
secondary coverts) and at the posterior of Concurrently, there is molt of the
the secondaries themselves. Both sexes scapulars and upper back. These new
have an extremely large alula, the func­ feathers are adult-like, being darker,
tion of which can only be theorized. broader, more firm, and comparable in
Both sexes have red bills and legs. The color to those adults or to immatures in
iris is said to be brown in both sexes advanced stages of plumage development.
(Delacour 1956) but specimens in the Specimens with 1 to 7 adult tail-feathers
Field Museum have labels indicating a uniformly show molt of the scapulars
white iris in adult males Merganetta a. and often have new feathers on the side
 Torrent Ducks 35

of the neck and on the head. Specimens Either this head molt precedes that of
with 4 to 10 adult tail-feathers also show northern races or too few specimens of
molt of the back, side, undertail-coverts, other subspecies have been collected at
rump, head and occasionally the chest. an appropriate time.
Belly molt is rare until the tail is fully Almost no data are available for
adult, but even then the belly feathers females and I suspect that many of the
are the last to be replaced, being preceded immature specimens are mislabeled.
by those on the chest, chest-side, and Many immature females have reddish
flank, in that sequence. on the chest but so also do males of Mer­
As Conover (1943) noted, the new ganetta a. armata. The best specimens
feathers which appear on juveniles are showing the transition from juvenile to
fully adult in color and pattern. There is nuptial plumage of females are F.M.
no positive evidence of a dull first non- 17432 (April—mostly juvenile) and inter­
nuptial plumage (the first basic plumage mediate specimens, which still have the
of Humphrey and Parkes 1959). There white belly but show a red throat (A.M.
are possible exceptions with regard to the N.H. 424855) or a red throat and chest
head and chest. The white facial area (F.M. 17433— taken in May). Several
becomes mottled white and black when near-adult specimens retain only a few
the distinct black pattern forms on males. white belly feathers (F.M. 17430—June;
Specimens of what are assumed to be F.M. No. 19202—August).
yearlings (see below) lack this mottling Most nesting of the southern form
so that it is lost either by molt or by probably occurs in October to January,
wear. The distinct pattern formed during with a peak in November. Johnson (1963)
this molt suggests that the latter may be reported two nests in October and
the case. Another area of uncertainty is November, and a female was captured
on the chests of both sexes when some on a nest in November at Cautín, Chile
rufous feathers appear during the loss of (F.M. 14296). Flightless juveniles have
the juvenile plumage. Possibly these are been taken in January and early March
first non-nuptial feathers but, if so, they (Tables I and II). Birds dominantly in
are neither widespread nor regular in all juvenile plumage were taken in January
individuals. to April, and birds dominantly in first-
Some specimens of the northern Mer­ nuptial plumage were taken from May
ganetta a. columbiana show little or no to September (Tables I and II).
head molt even when the tail is nearly Thus it appears that both sexes are
fully adult. However, several specimens essentially in full first nuptial plumage
of immature male Merganetta a. armata by the time of the normal breeding sea­
show the acquisition of its unique black son. Whether yearlings actually breed is
throat when there is no molt or little not known. There is no evidence of an
molt of the juvenile tail (Phil. 5442 and intermediate body molt in the spring in
F.M. 88246, respectively). New black yearlings, although tail replacement is not
feathers develop as a ‘V’ of two lines unlikely.
within the two halves of the lower man­ Once the first nuptial plumage has been
dible. Eventually a black throat and a acquired, the plumages of yearlings and
ventral mid-line on the neck are formed. adults are not readily distinguishable by

Table I. Series of specimens of male Merganetta a. armata showing the chronology

of acquisition of the first-nuptial plumage. For each column heading: + = some
present; — = none present.
Male
Juv. Nupt. Juv. Nupt.
Museum No. Date Flightless tail tail Body Body
A.M.N.H. 734385 January + + — + —
A.M.N.H. 734384 January + + — + —
F.M. 10400 March + + — + —
F.M. 88246 March — + + + —
R.O.M. 93534 May — — + + +
F.M. 14298 June — + + + +
F.M. 12263 June — + + + +
R.O.M. 93536 July — + + + +
F.M. 16097 July — + + + +
R.O.M. 93535 July — — + + +
F.M. 16096 September— — + + +
36 Wildfowl

Table II. Series of specimens of female Merganetta a. armata showing the chronology
of acquisition of the first nuptial plumage. For each column beading; + = some
present; — = none present.
Female
Juv. Nupt. Juv. Nupt.
Museum No. Date Flightless tail tail Body Body

A.M.N.H. 734383 January + + — + —

A.M.N.H. 424855 March — + + + +
F.M. 17432 April — + + + +
M.C.Z. 96607 April — — + + +
F.M. 17433 May — — + + +
F.M. 17430 June — — + + +
F.M. 19202 August — — + + +

body plumage. However, in examining amined. Ten of the 15 had tail molt
wing spurs and other characters as pos­ while only seven had body molt. Five
sible age criteria (see below), some of the seven in body molt were taken in
characters of the male wing were observed autumn (Feb.–April), one was taken in
which show reasonably good agreement spring (Nov.), and one in winter (June).
with known juveniles and with those in­ It seems probable that a late-summer or
dividuals assumed by their spur length autumn molt is regular, concurrent with
to be either yearlings or adults. As is and following the wing molt. More data
usual in waterfowl, the juvenile wing is are needed to determine whether a late-
retained for nearly a year until the simul­ winter or spring molt is regular in adults
taneous wing molt following the first but it seems common in other Neotropi­
breeding season. The greater secondary- cal anatids (Weller, in press).
coverts of the juvenile wing are white- Tail molt seems to be a more regular
tipped like those of adults but adult males and a less seasonally restricted event in this
have a slight black edge proximal to the species than in typical northern species.
white band, which is reduced or lacking Some tail molt was found in 79% of 49
in immatures. Females show a less clear­ adult birds, representing all seasons, ex­
cut pattern because both juvenile and amined for this feature. This suggests a
adult females have the black band on the constant replacement or a double molt
wing covens but usually it is larger than each year. There was no apparent serial
in adult males. pattern to the molt, although central tail­
feathers seemed to be replaced first.
Annual molt One female taken on a nest of 5 eggs
Following the acquisition of the first-nup­ (F.M. 14296) near Cautín, Chile, in
tial plumage, there is permanent sexual November, had tail molt but no body
dimorphism in color with no evidence molt. It did have long, plumose, gray
of any intermediate non-nuptial (‘win­ nest-down, characteristic of many nesting
ter’ or ‘eclipse’) plumage. Subsequent ducks (Weller 1957).
plumages are merely replacements of
feathers of the same color. Only three Wing spurs
flightless adults were observed in collec­ While examining the plumage of speci­
tions: a male from Junín, Peru (F.M. mens of various ages and both sexes, it
12797— no date), a male taken in Ecuador was noted that considerable variation
in April (F.M. 14268) and a female occurs in the size, shape and color of
taken in December in Ecuador (F.M. wing spurs. To clarify and quantify this,
11769). All three were in body molt but spur length, shape and color were re­
the new feathers were of the same color corded for 191 specimens.
as were the worn ones, confirming the Spur lengths were plotted according to
general observation that there is no season for males and females of all ages
‘eclipse’ plumage. and from all areas. Data on unsexed
Because of the extensive nesting period downy young were used as a base. A
of Torrent Ducks in northern breeding summary of means and ranges of speci­
areas (Johnsgard 1966), the chronology of mens from all regions is presented in
the plumage cycle is difficult to appraise. Table III, which shows clearly the pat­
Assuming that the breeding season is tern of size in relation to age. In addition
more restricted in the southern Mer­ 9 unsexed downy juveniles had the smal­
ganetta a. armata, six adult females and lest spurs (up to 1 mm). Because the
nine adult males of this race were ex­ great variation in the breeding season
 Torrent Ducks 37

tends to obscure patterns, data for Argen­ race, many birds still recognizable as sub­
tine and Chilean males only are plotted adults are present during June and a few
in Figure 1. are recognizable in October and Novem­
Growth of the spur starts as a rounded ber (Figure I). Thus, it appears that the
burr which is broader than long. Pre­ age of acquisition of the complete first-
sumably, growth of homy material in­ nuptial plumage is approximately nine
itiates at the base (Rand 1954) so that months to one year and that spur length
growth must be a continuous process that of males at that time is less than 6.5 mm
moves tissue upward to form a point in the southern race (Figure I), Spurs of
when the spur length equals its width. all such immatures are rounded or have
Spur size increases to 2.8 mm (both sexes a blunt tip. (Plate III a and b, facing
combined) when the juvenile is full­ page 44.)
winged. Thereafter rate of growth differs Spurs of adults develop an enlargement
by sex and when they are in nearly full at the base so that a constriction is
first-nuptial plumage males average 5.8 apparent near the point of attachment to
mm, females only 3.7 mm. From the the wing (compare Plate III b with c).
chronology of nesting in the southern Males in full-nuptial plumage have spurs

Table III. Spur length (mm) of Torrent Ducks of various ages and all sub-species.
Male Female
No. Av. Range No. Av. Range

Partial Juvenile 5 1.8 1.0–2.4 — — —

Flightless Juveniles 7 1.9 1.0–2.8 1 1.7 —
Juvenile 5 2.8 2.1–3.4 — — —
Immatures 6 3.8 3.2–3.8 2 3.2 3.0–3.4
Subadult 10 5.8 4.5-7.0 7 3.7 3.1–4.1
Adults 84 10.5 4.1-16.5 55 6.1 2.7–12.8

Figure 1. Spur lengths of males of Argentine and Chilean specimens plotted by age and
time.
38 Wildfowl

which tend to fall into one of the follow­ paring the age assessed from wing coverts
ing classes: short and blunt, intermediate with that derived from the spurs. Of 35
and pointed, or long and sharp (Plate III adult males checked, there was 89%
a, b and c). Long spurs often are ex­ agreement between age classification by
tremely sharp, occasionally are curved spurs and by wings, and agreement for
and may bear a translucent tip (c 15 young males was 93%. However, there
and d). Although the color of horny was only a 62% agreement for 21 adult
material varies with the conditions at the females. Only four young females were
time of preservation, blunt spurs are available and only one showed agreement.
dominantly yellow, the longer spurs horn- Problems of agreement in the adult males
brown or nearly black with a yellowish were about equally divided between inter­
tip. pretation of the spur length and of the
A frequency distribution (Figure 2) of wing characters. In females, wing charac­
spur length of males in full nuptial ters were the most variable.
plumage shows a small peak at six to No specimens were found without
seven millimeters, which may represent spurs, indicating that they are not shed.

Figure 2. Frequency distribution of spur lengths of juveniles, immatures and birds of both
sizes in adult plumage.

yearling males. The fact that some males Only one had been damaged in some way
in adult plumage have spurs no larger and this may have been by a shotgun
than birds still recognizable by their pellet. The function of these spurs is
plumage as subadults, further suggests still unknown, but they show no wear.
that males with spurs of less than 8 mm Occasionally, there are irregular rings
may be considered yearlings. The relative near the base or in the middle (Plate III
bluntness of the spur seems to be an even d) but they show no pattern suggesting
more reliable character. Blunt spurs, or annular growth rings.
spurs that are pointed but are as short To determine whether some internal
as they are wide, are probably those of characteristic of the spur might aid in
yearlings. Spurs which are longer than determination of age (as do annuli of fish
they are wide, with a well-developed scales or bones), spurs of three sizes were
point, presumably represent older males. sectioned by grinding. I am indebted to
Spur length probably increases with age Drs. Rand and Blake of the Field
but individual variation cannot be Museum for permitting the sectioning of
ignored. spurs from three specimens in their
Adult females have shorter spurs with charge. Three males were examined: a
less variation in shape. These also can be clearcut subadult (identifiable by some
grouped into categories, but there are white juvenile feathers on the abdomen),
many that are intermediate in length and
which cannot be aged with certainty. A which possessed a short, blunt spur (F.
frequency distribution of their lengths M. 208109); an adult with a sharply
shows no obvious breaks (Figure 2). pointed, black spur (F.M. 208112); and
A further check on the use of the spurs a male with the longest spur recorded (F.
as an age criterion was made by com­ M. 15682).
 Torrent Ducks 39

Figure 3. Spur size, shape and internal structure of a subadult male (Field Mus. 208109)
and two adult males (Field Mus. 208112 in center; Field Mus. 15682 at right).
Dots indicate points at which sections and measurements were made.

Cross-sections demonstrate that spurs core at the base of the spur, being bi­
have an outer horny layer, gray in adults, lobed in vertical cross-section (Figure 3)
gray or yellow in subadults. The major rather than a single spike and by being
portion of the spur consists of a clear mushroom-shaped at the base rather than
yellow material (presumably connective round. However, the significance of these
tissue) which surrounds and grows above differences cannot be appraised on the
a small whitish central bony core (Plate basis of this sample.
III e). A narrow dark-brown or black line
separates the outer gray horn and the Acknowledgments
central yellow connective tissue. The study was financed by a Chapman
Diagrams of these three spurs are Grant from the American Museum of
shown in Figure 3. They show the small Natural History. I am indebted to the
size of the true bony core. which appar­ following persons who made available
ently increases very little in length with collections in their charge and facilitated
age. They also demonstrate that much of my work: D. Amadon, E.R. Blake, J.
the growth of the spur is in the solid Bond, J. Barlow, K.C. Parkes, E.R.
horny tip. Although several irregularities Paynter, A.L. Rand, C.S. Sibley. My
of the horn layer in older birds suggests wife aided me materially in museum
new overlapping external layers, their studies and analyses. Mr. A.W. Johnson
pattern was not sufficiently distinctive to of Santiago, Chile, initially kindled my
be used as an age criterion. Both adults interest in the species and kindly took me
examined did have a more complex bony to an ideal observation area.

Summary
A study of nearly 200 museum skins of Torrent Ducks has made possible some generaliza­
tions on plumages and biology of this species, which will facilitate future field work.
Torrent ducks are distinctly and permanently sexually dimorphic, lacking any seasonal change
in coloration. Juveniles lack sexual dimorphism and are characterized by a white belly and
heavily barred sides. There is no evidence of a well-defined first non-nuptial plumage and
the first nuptial plumage seemingly is acquired by the time the bird is 9–12 months old.
Adults apparently have one complete annual molt in late summer but a partial spring molt
is probable.
40 Wildfowl

Both sexes have prominent metacarpal spurs on their wings. These spurs increase in
length, dark coloration and sharpness with age. However, most yearling males and some
females can be distinguished from adults by their short, blunt, yellowish spurs.
Spurs are made up of a small and short bony core, a more extensive layer of yellow con­
nective tissue, and a relatively thin outer layer of horn. No clearcut growth rings are
observable but the bony core, and probably the horn, increases in complexity with size.

References
Conover, B. 1943. A study of the Torrent Ducks. Field Mus. Nat. Hist., (Zool. series),
24:345–356.
Delacour, J. 1956. The Waterfowl of the World. Vol. 2. London: Country Life.
Humphrey, P.S. and K.C. parkes. 1959. An approach to the study of molts and plumages.
Auk, 76:1–31.
Johnsgard, P.A. 1966. The biology and relationships of the Torrent Duck. Wildfowl Trust
Ann. Rep., 17:66–74.
Johnson, A.W. 1963. Notes oo the distribution, reproduction and display of the Andean
Torrent Duck, Merganetta armata. Ibis, 105:114–116.
Phillips, J.C. 1926. A Natural History of the Ducks. Vol. 4. Boston: Houghton-Mifflin Co.
Rand, A.L. 1954. On the spurs on birds’ wings. Wilson Bull. ,66:127–134.
Sibley, C.C. 1957. The evolutionary and taxonomic significance of sexual dimorphism and
hybridization in birds. Condor, 59:166–191.
Weller, M.W. 1957. Growth, weights and plumages of the Redhead (Aythya americana).
Wilson Bull., 67: 189–193.

Prof. Milton W. Weller, Dept, of Zoology and Entomology, Iowa State University, Ames,
Iowa 50010, U.S.A.
Plate III. Wing spurs of Torrent Ducks Merganetta armata from Argentina and Chile.
(a) Juvenile male showing blunt point and wrinkled surface. (b) Yearling
male taken in September in nearly full first-nuptial plumage; spur point
becoming evident. (c) Bird presumed from its wing to be an adult at least
two years old. Note the elongated shape, acute and translucent tip (Scale
in cm). (d) Adult male showing overlapping layers of horn and translucent
tip. This was the longest spur recorded (16.5 mm). (e) Cross-section of a
spur near the base showing the central bone, the yellow connective tissue,
and the outer layer of horn. (See p. 33)
 WILDFOWL
19

Published by the Wildfowl Trust, Slimbridge

as a continuation of the Wildfowl Trust Annual Reports, 1–18

1968