CHAPTER 7

Impact of Sloths on N eotropical Forest

Energy Flow and Nutrient Cyclingl

G. G. MONTGOMERY AND M. E. SUNQUIST 2

Introduction

We report a field study of the role of two-toed and three-toed sloths (Cho-
loepus hoffmani and BradYPlIs infuseatus) in neotropical forest energy flow and
nutrient cycling. Sloths are arboreal herbivores with stomachs and digestive
processes showing convergence with the ruminants (reviewed by Goffart, 1971).
Sloth stomach anatomy and histology, which includes provision for long-term
retention of food, keratinous epithelium, and comparative lack of mucosal pa-
pillae, indicate that they should be considered as "bulk and roughage eaters"
(Hofman and Stewart, 1972), a category reserved in terrestrial ruminants for
grazers (grass feeders). Whether sloths are considered grazers on the basis of
their stomach characteristics, or browsers because they feed mainly (three-toed)
or partially (two-toed) on leaves, many of their effects on neotropical forest
processes would be expected to parallel the effects of terrestrial mammalian
grazers and browsers of grassland and savanna.
Annual net production of vegetation potentially available to browsers in moist
tropical forest ecosystems usually exceeds net production available to browsers
and grazers of tropical grassland and savanna (Golley, 1972). However, this
potentially large and diverse source of food for herbivorous tropical forest mam-
mals is not readily available because a large part of it grows in upper levels of
the forest (cf., Tadaki, 1966) and is out of the reach of terrestrial forms. There
are generally fewer mammalian herbivore species with terrestrial, as compared
with arboreal, life modes in tropical forest, relating to the lesser amount of food
available at the forest floor. Utilization of most of the vegetation available to
mammalian browsers in tropical forest requires adaptations both for moving

1 Dedicated to the memory of I. N. Healey, whose comments during our mutual stay
on Barro Colorado Island gave impetus to our work.
~ Present address: 4660 Heights Drive, Minneapolis, Minnesota 55421.

69

I!) by Springer-Verlag New York Inc., 1975

70 G. G. MONTGOMERY AND M. E. SUNQUIST

about in trees (Cloudsley-Thompson, 1972) and for digesting leaves. Require-

ments relating to these adaptations, such as limitations on maximum body size
(Bourliere, 1973), have apparently reduced mammalian radiation as arboreal
foliovores, thus there are relatively few mammalian species of arboreal foliovores
when compared with the great diversity of terrestrial mammalian herbivores
that evolved in tropical savanna and grassland. The great abundance and diversity
of potential herbivore food (especially leaves) might have provided for evolu-
tion of many more forms.
Among the mammals specialized for arboreal life, relatively few subsist solely
or primarily on leaves and buds. For example, 6 of 17 species of Australian
arboreal mammals and 4 of 44 species of Malayan arboreal mammals feed
primarily on leaves (Harrison, 1972). In general, however, those few arboreal
species that are foliovores live at densities that result in rather high species
biomass. Total mammalian herbivore biomass for tropical savanna and tropical
moist forest may be roughly equivalent (Eisenberg and McKay, 1973), yet the
number of species forming the herbivore biomass is much greater in savanna
than in moist forest (Bourliere, 1973).
No mammals are so well adapted for living in trees and feeding from them
as the sloths. In our attempts to understand the role of arboreal herbivores in
forest ecosystems, study of sloths would thus seem to be important. Prior to
this study, little was known of how sloths live in the wild and how they relate
to the forest systems in which they live. The one extensive field study (Beebe,
1926) did not emphasize the ecological relationships between sloths and plants.

Methods and Results

The role of sloths in the processes of energy transfer and nutrient cycling
involves (1) how they select their food from the array of plants available to
them; (2) the amount of food they crop from an area per unit time; and (3)
amounts of unused food they return as waste to the forest, where that waste is
returned, and rates at which it is recycled. Data on each of these topics will be
presented in turn, following the methods applicable to each topic. The data will
then be synthesized in the final section of the paper.
Since August 1970, 17 months of fieldwork have been completed on Barro
Colorado Island (BCI), Panama Canal Zone. Sloths were studied continuously
on BCI from August 1970 through September 1971 and periodically (J anuary-
March, August-September) in 1972. BCI is described in the work of Standley
(1933), Carpenter (1934), Hladik and Hladik (1969), and others.
The approximately 15 km 2 island was formed in 1911-1914 with the flood-
ing of Gatun Lake, which forms a fresh-water portion of the Panama Canal.
The ship channel passes to the east and north of BCI, and one or more km of
open water separate the mainland and the island there. However, even this ex-
panse of water does not isolate the BCI sloth population from the mainland
 Sloth Impact on Energy and Nutrients 71

population. Sloths swim well (Beebe, 1926; Worman, 1946) and a three-toed
female swam across the ship channel and set up residence on BCI (Montgomery
and Sunquist, unpublished). Furthermore, for about one-fourth of its circum-
ference on the southwest side, the island is separated from the mainland by a
much narrower channel which sloths could easily swim and thus move between
BCI and the mainland.
The forest on the portion of the island where our study sloths lived (Ben-
nett, 1963) is of young to medium age with a lower stratum reaching to about
40 ft (12 m) and an upper stratum reaching usually to 80-100 ft (24-30 m).
Parts of the area have been recently disturbed, which, in combination with tree-
falls that opened small areas of canopy, resulted in mixing of forest age and
vegetation. Perhaps as many as 800 species of plants were available to the sloths
as potential food and as support for their movement and other activities.

Capture, Radio Marking, and Radio Location of Sloths

Radio location of free-living sloths provided much of the data with which
we determined home range characteristics, movements, and habitat. Sloths were
captured for radio marking by placing a noose around the animal while it was
in the tree. The animal was then noosed, pulled back to the climber, placed in a
canvas bag, and lowered to the forest floor.
When a sloth could not be captured with the noose, branches were sawed
until the sloth fell to the ground with the foliage. After capture, three-toed sloths
were easily handled and marked. However, it was necessary to anesthetize two-
toed sloths with ether or Halothane (Fluothane; Ayerst Lab. Inc., N.Y.).
Fifty-two three-toed and 12 two-toed sloths were captured. Of these, 31
three-toed and II two-toed sloths were radio marked and tracked. Some indi-
viduals were handled only once; most were recaptured at least once so that we
could change transmitter batteries or examine the animal. One three-toed sloth
was captured I I times. Nine adult three-toed sloths and six adult two-toed sloths
were radio tracked for periods of three or more months and provided most of
the data for this report.
Sloths were radio marked with transmitters (Model ST -I and MT -1; A YM
Jnst. Co., Champaign, 111.), which were packaged as neck collars. Transmitter
packages for adults weighed about 50 g, and those for juveniles 20 g or less.
Each transmitter broadcast at two exclusive frequencies (ncar 150 megahertz
and near 450 megahertz), and individual sloths were identifiable by different
such pairs of frequencies. We rarely saw the animals, and usually identified them
by these frequencies. After a sloth was marked, examined, weighed to the near-
est 0.1 Ib, and standard body measurements taken, it was released at the base
of the tree where captured.
We usually radio located (Montgomery et aI., 1973) each animal once per
day. Series of hourly locations made for 24-hr periods confirmed that a single
daily radio location was adequate to locate a three-toed sloth for that date;
they seldom left a tree and then returned to it on the same date. However, two-
 72 G. G. MONTGOMERY AND M. E. SUNQUIST

toed sloths moved from tree to tree at night (Sunquist and Montgomery, 1973)
and probably used more trees than the single daily radio locations indicated.
When possible, each radio location was visually confirmed.
Each tree used by a sloth was numbered and marked. Later the tree location
and elevation at its base were determined. The species of the tree was identified,
and its DBH, height, and crown dimensions were measured. Percentage of the
crown exposed to sunlight, relative amount of vines in the crown, and relative
number of pathways into the crown also were estimated.

Home Ranges, Movements, and Habitat Use

The home range of a sloth of either species is best characterized as a set of

tree crowns; a part of each crown is used for one or more days. A sloth moves
among these trees via interconnecting tree crowns and vines. Places on the
forest floor to which the animal descends periodically to defecate are included
in a home range, as are pathways from the canopy to the forest floor. Vines and
the boles of smaller trees are employed by the animal during descent.
An idealized top view of the home range of an adult female three-toed sloth
(Figure 7-1) illustrates some features of sloth home ranges. This range is based
on 187 radio locations during one year. In this period the animal was located
in 63 trees of 35 species. About half of these were in the forest canopy; others

Three·toed Sloth No. 2, Adult Female

187 locations, August 1970 to September 1971
63 Trees, 35 Species
~-
8~
Key
{0 8

8;
/
Canopy Tree
~ Understory Tree /'"
Crown
Diameter

(M)
o 30 61
o 100 200
Scale (FT)

.~

Fig. 7-1. Idealized top view of the home range space of an adult female three-
toed sloth. Canopy trees were those that had 80 percent of the crown or more exposed
to direct sunlight. Tree crowns were drawn on the basis of crown width; not all crowns
were circular. Black dots show the location of the hole of each tree.
 Sloth Impact on Energy and Nutrients 73

were in the upper understory. All 63 trees were in an area 250 by 500 ft (76
by 152 m). Thus the area of the home range, based on the minimum-size rectan-
gle that enclosed all trees, was slightly more than 1 ha.
Sloths are not sedentary, but regularly move from tree to tree (Sunquist and
Montgomery, 1973) . Most two-toed sloths changed trees on successive dates,
while three-toed sloths wcre in the same tree on successive dates about 40 percent
of the time. Three-toed sloths moved less than 125 ft (38 m) per day in 89
percent of the cases, whereas two-tocd sloths moved more than 125 ft (38 m)
per day in 54 perccnt of applicable measures .
Changes in location by a sloth have vertical as well as horizontal compo-
nents. For example, an adult femal e three-toed sloth changed trees 18 times,
using 13 trees of 10 species over a period of 33 locations (Figure 7-2). She
moved vertically within thc range of 25 to 80 ft (7.6 to 24.4 m) above the
forest floor (Figurc 7-2) . Except when defecating, sloths were usually more
than 30 ft (9. 1m) above th e forest floor and were at times near the tops of the
tallest trecs. Sloths of both species werc located 80 ft (24.4 m) or more above
thc ground more than 20 percent of the time (Tablc 7-1) .

no. 2, Adult Female

33 Dates, 29 Locations
13 Trees, 10 Species
18 Moves

IMI
oj
30 61
o 100 200
HOrizontal Scale (FT)

HOrizontal Movement

_ 100_ _ 61
l-
Ll..
•
.<:
.~ 50- ••• • • •• - 30~
I •••••••••••• • • • •••••
O- ~,
•
__________ __________
~
, ~,__________ ~, -J _O

20 30 10 20
August September
Vertical Movement

Fig. 7-2. Hor izontal and vertical movements of an adult female three~toed sloth.
Arrows indicate the direction of movement during her moves from tree to tree. Black
dots for horizontal movements indicate the locations of the boles of trees she used.
74 G. G. MONTGOMERY AND M. E. SUNQUIST

Table 7-1. Home range sizes, number of trees used, and number of tree species used
by radio-marked sloths. a

Sloth Area of Home Range

Species (ha) Number of Trees Number of Tree Species

Two-toed 1.6 28 18
0.4 14 12
3.9 27 17
3.3 34 17
1.2 19 16
1.4 26 16
Three-toed 1.1 63 32
2.8 59 24
1.8 33 19
0.5 37 28
0.8 33 20
1.8 41 21
1.1 27 21
0.7 38 25
3.7 42 30

• Data are shown for six adult two-toed sloths for which we had three or more months
of data, and for nine adult three-toed sloths for which we had five or more months of data.
Size of each home range was estimated by' the minimum-size rectangle that enclosed all trees
in which the sloth was radio located.

Trees used by a sloth of either species were within a relatively small area.
The six two-toed sloths (Table 7-2) had an average home range size of 1.97 ha.
The average home range size for nine adult three-toed sloths was 1.60 ha. Total
area is probably not an appropriate parameter for characterizing sloth home
ranges. Because trees within an area were selectively used, the home range is
better typified by the number and species of trees used by an animal. The six
two-toed sloths (Table 7-2) used an average of 24.7 trees each and used a con-
sistent number of tree species, averaging 16 species each. The nine three-toed
sloths used an average of 41.4 trees each; they averaged 24 species per individual.
Approximately 53 percent of all trees used by sloths of either species were
used only once by that sloth (Figure 7-3). Thus a few trees accounted for most
of an individual's radio locations. Eight or fewer trees accounted for 50 percent
or more of the radio locations of each of the six adult two-toed and each of the
nine adult three-toed sloths.
Likewise, a few species of trees accounted for most of an individual's radio
locations. Approximately 40 percent of all species used by individuals were used
only once (Figure 7-4). For two-toed sloths, four or fewer species of trees
accounted for 50 percent or more of the radio locations of each of six animals.
Eight or fewer species accounted for 50 percent or more of the locations of each
of nine three-toed sloths.
Each sloth had a modal tree in which it was radio located most often (Table
7-2); modal trees were seldom used by other sloths. No two sloths had the same
 Sloth Impact on Energy and Nutrients 75

Table 7-2. Modal trees and tree species of six adult two-toed sloths and of 9 adult
three-toed sloths that were radio located over periods of three months or more.

Percent Percent
Modal Total In in
Animal Tree Loca- Modal Modal
Number Number tions Tree Modal Tree Species Species
- -- ------ .. - ----- --- --------- - -------- - - - - - - - - -

Two-toed
3 109 47 19.1 A nacardilllll excelsul1l 25.5
4 185 95 57.8 A Ilacardilllll excelslIl1l 70.5
5 288 50 12.0 A llacllrdiwl1 excelslIlIl 16.0
6 444 69 11.6 Anacardilllll excelslIlIl 30.4
7 66 37 16.7 Dipteryx panalllensis 16.2
8 426 58 13.8 Spolldias nigrescens 17.2

Three-toed
202 71 186 9.1 GlIstavia sllperba 11.5
205 138 163 13.5 A Ilacardillm excelslllll 15.1
206 18 196 21.4 T etragastris pana/1lensis 28.6
208 103 214 20.6 F iClI.I' trigonata 20.6
219 490 110 14.5 Ficlls obwsifolia 14.5
221 84 118 10.2 A nlicardilllll excelsllm 14.4
225 379 83 15.7 lnga sp. 15.7
227 358 131 13.0 POlIlsenia armata 17.5
231 205 93 10.7 Dipteryx pallalllensis 10.7
--------- ----

modal tree, although the home ranges of many animals overlapped. A two-toed
sloth was in its modal tree more than 10 percent of the times it was located, one
animal being there about 60 percent of the times. A three-toed sloth used its
modal tree an average of 14 percent of the time.
Each sloth also had a modal tree species (Table 7-2) that in some cases
was different from the species of the modal tree. Four of six two-toed sloths used
Anacardium excelsum most often. Three-toed sloths were less consistent. Two
of nine were found most often in Anacardium excelsum. The remaining seven
animals used seven different modal species.
Sloths used a wide variety of trees (Table 7-3) even though the distances
they moved were relatively short. In total, 106 tree species were used by sloths.
Three-toed sloths were radio located in 98 tree species, and two-toed sloths in
52 tree species. There was a good deal of overlap in tree species, with two-toed
sloths using 41.5 percent of the species used by three-toed sloths and three-toed
sloths using 84.6 percent of the species used by two-toed sloths.
As an example of the complex sequence of use of various species by sloths,
an adult female three-toed sloth (Figure 7-5) regularly left and returned to
Gustavia superba, staying an average of 1.6 days per visit to that species. Cecro-
pia eximia trees were visited twice, while the remaining eight species were vis-
ited only once during the 33-day period. Distances the animal moved among
these species were small, ranging from 17 to 372ft (5.2 to 113.4 m), as meas-
ured in three dimensions.
76 G. G. MONTGOMERY AND M. E. SUNQUIST

6 Adult Two-toed
60

40

.c
20
0
en
.c
u
'"
UJ
>
D 0-
.,
"C
10 15>15
.,.,.
:5
~
80
....«
...c0 9 Adult Three-toed
Q)

.,~ 60
0- 9

40

Locations in Each Tree

Fig. 7-3. Frequency distribution of the percentage of all trees used by each two-
toed sloth (top) and each three-toed sloth (bottom) that were used once, twice .... N
times by the individual. Black dot indicates the mean for the number of animals shown
above the line, which indicates the range.

Three-Toed Sloth Food Habits

We rarely observed three-toed sloths feeding; two-toed sloths are nocturnal

(Sunquist and Montgomery, 1973) and were never seen feeding. Samples of
stomach contents from three-toed sloths were collected with a tube inserted
orally through the esophagus (Montgomery, 1969). We were unable to collect
 Sloth Impact on Energy and Nutrients 77

80-

6 Adult Two-toed
60

40
.r;
.2en
..c
u
20
W'"
>
..c
""C

:s
OJ
0
I I
on
OJ 5 15>15
'uOJ
c.
en
OJ
OJ
80
~
«=
'0
...c 60
9 Adult Three-toed
OJ
~ 9
OJ
0...

40

20

5
Locations in Each Tree Species

Fig. 7-4. Frequency distribution of the percentage of all tree species used by each
two-toed sloth (top) and each three-toed sloth (bottom) that were used once, twice,
... N times by the individual. Black dot indicates the mean for the number of ani-
mals shown above the line, which indicates the range.

such samples from two-toed sloths. Portions of the samples were prepared by
methods adapted from Stewart (1967) and compared histologically with leaf
material from known species of trees and vines. Identification of unknown leaf
material was made on the basis of shape and size of the stomata and other
epidermal characteristics (Baumgartner and Martin, 1939). The collection of
known leaf material was incomplete; thus many species in samples remained
unidentified.
 Table 7-3. Tree species used by two-toed and three-toed sloths and other arboreal fauna. -..J
00

Two-toed Three-toed Three-toed Howler Spider Cebus

Tree Species N % N % Sloth Monkey Monkey Monkey
- ------- ----

Acanthaceae
Trichanthera gigantea 3 0.8
Anacardiaceae p
Anacardium excelsum 117 30.9 92 4.8 0 C
Astronium graveolens 5 1.3 P
Mangi/era indica 1 0.3 I 0.05 ~
Spondias mom bin 1 0.3 3 0.1 0 C 0
Spondias nigrescens 17 4.5 45 2.3 ..,z
Cl
Annonaceae 0
Guatteria dumetorum 8 0.4 ~
ttl
.Apocynaceae :<l
><:
Lacmellea panamensis 0.3 131 6.8 a,s >
Araliaceae Z
0
Dendropanax arboreus 0.05
Didymopanax morototoni 0.05 C ~
Bambusa tTJ
Bamboo sp. 0.05
CIl
Bignoniaceae c::
Jacaranda copaia 19 1.0 0 z
,0
Tabebuia rosea 38 2.0 c::
Bombacaceae ..,Vi
Bombacopsis quinata II 0.6 C
Bombacopsis sessilis 15 4.0 13 0.7 S
Cavanillesia platani/olia 2 0.5
Ceiba penlandra 5 1.3 30 1.5 0 H
Ochroma pyramidale 2 0.1
Pseudobombax septenalum 30 1.5 0
Quararibea aSlerolepis 0.3 4 0.2 H H
Boraginaccae
Cordia alliodora 0.3 9 0.5
Burseraceae
Protium costuricense 14 0.7
Protium sp. 1 0.3 17 0.9
Protium telluijo/iunJ 4 1.1 26 1.3 0 C
Tetragastris panamensis 1 0.3 79 4.1 0
Trattinickia aspera 11 2.9
Combretaceae r:/).

Terminalia amazonica 5 1.3 0.05 0 0"

.....
0'"
Elaeocarpaccac
H
Slonea terniffora 1 0.3 2 0.1 a
Euphorhiaceac "d
$ll
Acalypha diversi/olia 1 0.05 ~
A lchornea costaricensis 2 0.5 24 1.2 0
;:l
Croton hillbergianus 1 0.05
H ura ('rcpitans 4 0.2 tn
;:l
('1)
Hyeronima laxif/ora 8 2.1 38 2.0 .....
OQ
SapiulIl caudatwll 4 1.1 6 0.3 '<
FI acourtiaceae $ll
;:l
Casearia arborea 2 0.1 0..
Tetrahy/aciulIl johansenii 3 0.1 Z
Zue/ania guidollia II 0.6 ~
....
Guttiferae (i)'
Ca/op/zylllllll /ongifo/iulIl 0.3 25 1.3 C ~
[J>

Rheedia lIladruno 10 0.5

Symphonia g/obulijera 8 0.4
Lauraceae
Bei/schmiedia pendu/a 0.3 9 0.5
Nectandra sp. 5 0.2
Lccythidaceae
COllratari panalllensis 0.05
-.]
Gustavia superha 3 0.8 39 2.0 S \0
 Table 7-3. (Continued) 00
0
- -----------------

Two-toed Three-toed Three-toed Howler Spider Cebus

Tree Species N % N % Sloth Monkey Monkey Monkey
---~.

Leguminosae
Andira inermis 2 0.1 a
Dipteryx panamensis 15 4.0 73 3.8 S
Inga goldmanii 2 0.5 47 2.4
P
Inga marginata 1 0.05 P
Inga guaternata 1 0.05
~
Inga sp. 37 1.9 a 0
Lonchocarpus sp. 4 0.2 ...,z
Platymiscium polystachyum 3 0.8 7 0.4 C Cl
0
Platypodium elegans 1 0.3 10 0.5 ~
ttl
Prioria copaijera 10 2.6 2 0.1 :;0
Pterocarpus hayesii 1 0.3 ><
Meliaceae >
Z
Guarea guidonia 2 0.1 a I:)

Trichilia cipo 0.3 79 4.1 a ~

Moraceae
Artocarpus communis 0.05 rn
Brosium bernadetteae 3 0.8 til
Cecropia eximia 146 7.6 a,s H H c:
z
I:)
Cecropia obtusijolia 14 0.7 a,s
Coussapoa panamensis 2 0.1
c:
en
...,
Ficus costaricana 1 0.05 C
Ficus insipida 4 1.1 11 0.6 H
Ficus obtusijolia 1 0.3 21 1.1
Ficus popenoei 3 0.1
Ficus sp. 20 1.0 a
Ficus tonduzii 3 0.1 C
Ficus trigonata 62 3.2 a
Ficus yoponensis 0.3 20 1.0 a
 MagI/ira costaricana 10 0.5
Poulsenia armata 10 2.6 114 5.9 0 H H
Pourouma aspera 2 0.1
Trophis racemosa 6 0.3
M yristicaceae
Virola nohilis 13 3.4 49 2.5 0
Virola sebi/era 4 1.1 28 1.4 0
Myrtaceae
Eugenia nesiotica 1 0.3 1 0.05 CIl
Eugenia sp. 2 0.1 0'
....
N yctaginaceae P"'
G uapira stllndleYllllum 2 0.1 ......
Olacaceae S
'0
~
Heisteria concinna 9 2.4 3 0.1 ~
Palmae 0
Astrocaryum standleyanum 10 0.5 H ::s
Scheelia zonesis 1 0.05 tI1
Rhizophoraceae ::s
(D
..,
(Jq
Cassipourell elliptica 2 0.1
'<
Rosaceae
~
Hirtella americana 2 0.1 ::s
0-
Hirtella sp. 3 0.1 0
Hirtella triandra 3 0.1 Z
s=
Licania Izypoleucll 5 0.2 ..,....
~.
Licania platypus 7 1.8 26 1.3 0 ::s
Rubiaceae
....
til

A lseis blackiana 7 1.8 27 1.4

Genipa americana 10 2.6 6 0.3
Guettarda foliacea 1 0.05
Macrocnemum glabrescens 5 1.3 30 1.5
Posoguerill latifolia 2 0.1
Rutaceae
Zanthoxylum panamense 1 0.3 6 0.3
Zanthoxylum procerum 1 0.3 00
.....
 Table 7-3. (Continued) 00
N

Two-toed Three-toed Three-toed Howler Spider Cebus

Tree Species N % N % Sloth Monkey Monkey Monkey

Sapindaceae
Cupania sylvatica 0.05
Sapotaceae a
Chrysophyllum cain ito 1 0.05 C
Chrysophyllum panamense 7 1.8 33 1.7 P
Pouteria stipitata 4 1.1 20 1.0
Solanaceae ~
0
Solanum hayesii 3 0.8 ..,Z
Sterculiaceae C!
0
Guazuma ulmi/olia 1 0.05 ~
Sterculia apetala 4 0.2 ttl
0.3 :;<l
Tiliaceae ><
Apeiba membranacea 15 4.0 74 3.8 0 >
Apeiba tibourbou 14 0.7
z
0
Luehea seemannii 19 5.0 40 2.1 0
Ulmaceae ~
Trema micrantha 0.05 rn
CIl
Evidence for use of leaves as food: c::
Z
S = stomach contents, this study 10
o = observed; this study c::
H = observed; Hladik and Hladik (1969) Vi
..,
C = observed; Carpenter (1939)
 Sloth Impact on Energy and Nutrients 83

Magui~a .. _ _8~9 ____ Cecropia Luehea _ _ ~5~ _ - . Sapium

;:
costaflcana exemta seemannii cauda tum

113.4

104.6 ~ Gustavia 34_._7_... Poulsenia

Macrocnemum _ _ _
superba glabrescens. armata
34.1

Genipa Anacardium •• _ _....;2;".6_.8_ _ __ Astrocaryum

americana excelsum standleyanum

Fig. 7-5. Sequential use of various tree species by an adult female three-toed sloth
during a 33-day period. Distance in meters that the sloth moved between trees of
the various species were measured in three dimensions. Sequence of use of the tree
species is indicated by arrows; dashed lines indicate missing data.

Twenty-eight tree species and three vine species were eaten by three-toed
sloths (Table 7-3). We observed three-toed sloths feeding on the leaves of 25
different tree species and one vine species. Three additional tree species and
two vine species were identified from three-toed sloth stomach contents.
Our finding that sloths fed on at least 31 plant species and probably many
additional ones is in contrast to the conclusion (Beebe, 1926) that three-toed
sloths readily accept only leaves from Cecropia trees, but will eat Spondias
leaves when Cecropia is not available. Lundy (1952) also felt that three-toed
sloths lived almost exclusively on Cecropia leaves. Others have observed three-
toed sloths feeding on several plant species (reviewed by Goffart, 1971), in-
cluding leaves from the genera Ficus, Spondias, Eriobotyra, Luehea, Bombax,
Chorisia, and Clusia. Not included in Goffart's review was the observation of
sloths feeding on additional species of the genera Hevea, Elizabetha, and Ceiba
by Carvalho (1960), who noted that Cecropia trees were abundant but that three-
toed sloths were not often found in them.
In spite of these reports, the idea that three-toed sloths prefer Cecropia, an
idea not originally based on systematic field study, is still strongly entrenched
(cf. Goffart, 1971). Although some study animals were never radio located in
Cecropia trees, our data are not complete enough to completely refute the idea
that sloths are dependent on Cecropia. They do, however, certainly suggest major
roles for many other tree species in three-toed sloth food habits.
Sloths fed on and used some of the same species of trees that were used by
primates on BCI (Hladik and Hladik, 1969; Carpenter, 1934). Of the 28 tree
84 G. G. MONTGOMERY AND M. E. SUNQUIST

species known to have been used by three-toed sloths as food, six were used as
food by primates. Of the 98 tree species in which three-toed sloths were radio
located, 16 were fed on by primates.
Hladik and Hladik (1969) estimated that 50 percent of the diet of howler
monkeys (Alouatta), 30 percent of that of night monkeys (Aotus), 20 percent
of that of spider monkeys (Ateles) , and maximums of 15 and 10 percent of
those of Cebus and Saguinus, respectively, were leaves; thus there is potential
competition for food between primates and sloths. On numerous occasions we
radio located sloths in trees that were occupied also by howler or spider mon-
keys, and saw no behavioral manifestations of this competition. A more impor-
tant point for the forest, however, is that the effects of cropping by sloths,
primates, and perhaps iguanas (Iguana iguana), as well as by invertebrate herbi-
vores, are additive.
A home range was not exclusively occupied by one sloth, but was shared with
several other sloths of both species. As an example of the overlap of home
ranges, 22 of the 63 trees used by an adult female three-toed sloth were also
used by at least one other radio-marked sloth, none of them at the same time as
the female. A total of nine different radio-marked three-toed sloths and four
different radio-marked two-toed sloths used portions of the female's home range.
Possibly other sloths shared her home range as well.
Adult sloths of both species tend to be solitary, and we very rarely saw two
or more of the same species in the same tree. Beebe (1926) noted that three-
toed sloths were solitary, and described antagonism that occurred when two or
more were placed in the same cage. Two-toed sloths show more tolerance and
are kept in groups of two or more per cage (Veselovsky, 1966; McCrane, 1966;
Van Doorn, 1971). There were several instances when we observed a two-toed
and a three-toed sloth in the same tree at the same time, but because their activ-
ity patterns are different (Sunquist and Montgomery, 1973), they were probably
in the tree for different reasons. The two-toed sloths rested during the daytime
when radio locations were made, while the three-toed sloths fed, groomed,
rested, and carried out other activities.
Because of the overlap in home ranges, several sloths often visited the same
tree, but as a rule only one animal was a frequent visitor to the tree, other indi-
viduals visiting it for one to two days at most during the study. Some trees were
much visited during the study, such as a small Lacmellea panamensis, which was
known to have been used by four sloths for a total of 87 days.

Leaf Production

Estimates of the proportion of the total leaf production used by sloths were
based on two areas (0.336 and 0.434 ha). Three-toed sloths were located on
the sample areas 438 times, while two-toed sloths were located there 49 times.
In these areas all trees with more than 0.5 ft (15.2 g) DBH were identified and
measured. In this area there were an average of 358 trees per hectare. Of these
trees, 76 were used by one or more sloths (Table 7-4).
 Table 7-4. Summary of biomass and cropping rate estimates for population classcs of two-toed and three-toed sloths."
en
Two-toed Sloths Three-toed Sloths 0"
Adult Juvenile Ad. Male Ad. Female Juvenile Young So
......
i3
Total Sloths Captured 10 2 17 18 6 11 "d
$>:l
Percent of Species Population 94.3 5.7 32.7 34.6 11.5 21.1 n
,.,.
Average weight (Kg) h 3.50 ± .79 l.05 ='= .10 2.78 ± .42 3.18:1: .74 1.06 _1= .50 0.73" a
Individuals/ha 0.99 0.06 2.77 2.93 0.97 1.79 ::s
Biomass/ha (Kg)h 3.46 0.06 tI1
::s
(1)
Percent of Species Biomass 98.2 1.8 7.70 9.32 1.03 1.31 "1
Percent of Total Sloth Biomass 15. I 0.3 39.8 48.1 5.3 6.8 ~
Annual Cropping Rate 33.6 40.7 4.5 5.7 $>:l
(Kg of leaf per ha) 4.29 0.04'1 14.24 17.24 0.95 11 1.21 d ::s
p,.

a Sloth weights were averaged (95 percent confidence intervals are shown) from weights of all animals captured during the study. except that
Z
~
,.,.
weights of young three-toed sloths were estimated on the basis of average weight when they became independent of their mothers. No young "1
(ii'
two-toed sloths were captured. See text for explanation of methods for calculating the other entries in the table. ::s
h Weights are corrected for gut contents by gross weight -30 percent.
,.,.
en
,. All young assumed to weigh 1050 g, the average weight at social weaning (Montomery and Sunqllist in ms.).
" Assllmmed cropping rate is Y2 of adult cropping rate.

00
Ul
86 G. G. MONTGOMERY AND M. E. SUNQUIST

Leaf production for each tree was assumed to be proportional to the cross-
sectional area of the tree crown, computed by multiplying the measured width and
depth of the crown. The sum of the cross-sectional areas of the crowns was
379,597 sq ft (35.255 m 2 ) per hectare. The sum of the crown area used by
sloths was 40 percent of the total for all trees.
Our best estimate of the amount of leaf material produced annually per
hectare on BCI comes from the leaves that fall to the forest floor. This estimate
is lower than the amount of leaf available to sloths, because cropping occurs
before the leaves fall. Foster (unpublished) estimated annual leaf production
at about 6,000 kg/ha on the basis of leaf-fall traps samples. Ovington and Olson
(1970) estimated an average leaf biomass in Puerto Rican montane forest at
7,880 kg/ha. Odum et al. (1970) gave an estimate of 8,120 kg/ha for lower
montane forest in Puerto Rico. Odum (1970) also used an estimated average of
9,390 kg/ha in modeling the Puerto Rican forest.
Sloths of both species descend to the forest floor to defecate. Feces are accu-
mulated by an animal for a period of time and passed periodically on the forest
floor as small, hard pellets. Defecation rates were used as a basis for censusing
sloths, while weights of fecal masses were used for estimates of amounts of leaf
processed by sloths per unit time. The defecation rate of three-toed sloths was
estimated in a field experiment involving five adult animals. All animals were
radio located daily during the experiment.
Each animal was recaptured at the beginning of the experiment and again
within two days after it defecated. A radio transmitter, measuring about 5 by
15 mm and weighing about 3 g, was inserted into the rectum through the anus.
Transmitters were about twice as large as fecal pellets produced by the sloths.
The transmitter was passed with the feces when the sloth defecated, and allowed
radio location of the fecal mass. Estimates of defecation rate were based on
intervals, to the nearest day, between successive defecations. Feces were cleaned
in tap water and dried in an oven at 80 C for two weeks before the mass was
0

weighed to the nearest 0.01 g.

In this study all feces were deposited within 10ft of a tree used by the sloth.
Three-toed sloths buried their feces by punching a depression in the leaf litter
or soil with their tail. Two-toed sloths, which lack a tail, usually left their fecal
masses on the soil or litter surface. Feces of two-toed sloths were distinguishable
from that of three-toed sloths on the basis of size and shape of the fecal pellets.
Thirteen intervals between successive defecations were determined for five
adult three-toed sloths. Based on these intervals, the sloths defecated on the
average once every eight days (±2 days; p < 0.05). Dry weights of fecal masses
ranged from 45 to 68 g. The mean dry weight was 56 g (± 14 g; p < 0.05).
The mean amount of feces accumulated per sloth per day was 7.4 g (± 1.6 g;
p < 0.05).
Assuming a "digestive efficiency" (Kleiber, 1961, p. 253) of 50 percent,
each sloth cropped an average of 14.7 g of leaves (dry weight). The assumed
efficiency is similar to that for tropical ruminants fed on leaves (Hussain et aI.,
1960a, 1960b). We avoided many of the sources of error that can occur when
food intake rates are estimated on the basis of feces weight (Galley, 1967, pp.
 Sloth Impact on Energy and Nutrients 87

111-115) by using free-living sloths in the experiment. The normal feeding

pattern of the animals was probably not changed significantly, either in choice
of diet or amounts cropped.
Sloths used in the experiment ranged in weight from 3.63 to 4.99 kg (mean
4.16 kg ± 0.96 kg; p < 0.05). Goffart (1971) suggested a downward cor-
rection of 30 percent of the body weight to account for the weight of the gut
contents. Biomass estimates and sloth body weights used throughout this paper
were calculated on the basis of gross body weight minus 30 percent. After cor-
rection, sloths in the experiment had a mean body weight of 2.91 kg.
The three-toed sloths accumulated and passed feces at an average rate of
2.5 g/kg of sloth per day. This represents an estimated cropping rate of 5.1 g
of leaf per kg of sloth per day, using an assumed digestive efficiency of 50 percent.
With the best available published information, we estimate that howler mon-
keys may crop leaves at a rate (38 g/kg per day) more than seven times the
three-toed sloth cropping rate. This rough estimate is based on visually estimated
quantities of fresh leaves consumed during daily periods in two examples by an
adult monkey (Hladik and Hladik, 1969) converted to dry weight by using an
estimated leaf water content of 60 percent (Lieth, 1970), and assuming an adult
howler monkey body weight of 5.5 kg (Eisenberg and Thorington, 1973).
Three-toed sloths cropped leaves at a rate one-fourth or less the intake rates
of terrestrial mammalian grazers in Africa (reviewed by Petersen and Casebeer,
1971). As examples, Grant's and Thompson's gazelles (Gazella granti and
G. thol1lsollii) crop at rates of 22.2 and 26.5 g/kg body weight per day, while
hartebeests (A lcelaphlls bllselaphus cokii) consume 19.9 to 3 1.8 g/kg body
weight per day.
The lower intake rate we found for three-toed sloths is probably related to
( 1) their lower basal metabolic rate, which is 51 percent lower than those of
mammals of comparative size (Goffart, 1971), and (2) the daily lowering of
their body temperature with reduction in ambient temperature and available
sunlight (Montgomery and Sunquist, unpublished). The combination lowers
their total metabolic demand, although their total daily activity is comparable
with that of other mammals; they are active for about 10 hr of each diel (Sun-
quist and Montgomery, 1973). The lower individual metabolic demand and
related reduced intake of food, in comparison with those of more metabolically
active animals such as the howler monkey, result in a reduced individual effect
on the forest and would allow maintenance of higher sloth population densities
with the same total effect on the food supply. Retention or reevolution of the
primative trait of poikilothermy might thus be a part of a life history strategy
that allows three-toed sloths to remain cryptic, relatively sedentary forms while
maximizing opportunities for contact and mating.

Census of Sloths

The census was based on counts of the fecal masses deposited by sloths in an
area of known size during a known time period. The average defecation rate was
88 G. G. MONTGOMERY AND M. E. SUNQUIST

then used to estimate the number of individuals in the area during the census
period. This method involved many of the sources of error inherent in pellet-
group count censuses of ungulates (Overton and Davis, 1969). These sources
are considered below.
Sloth feces decompose very slowly over several months, and differential de-
composition during the period should have been negligible. Dung beetle or other
animal activity was not important. The influence of rain also was not important
since the census was conducted during the dry season. Some fecal masses may
have been missed, but bias in successive searches should have been equal. Dis-
turbance of normal defecation patterns of the animals was probably minimal.
The sloths were many meters above our heads while we searched the area and
were not obviously disturbed by our presence. Radio-marked animals continued
to use the census area during and after the searches.
An area of 0.324 ha was selected for census. We twice searched the entire
area after moving leaves and other material aside with rakes; each search re-
quired more than a week of effort. The average interval between searches was
35 days (27 January to 2 March 1972). Leaves were replaced to minimize dis-
turbance of the forest floor. When fecal masses were found, they were marked
and covered with nylon netting. Thus we could distinguish masses found during
the first search from those placed in the census area between searches, and to
examine fecal masses six months and more after their initial discovery.
A total of 14 fecal masses (twelve three-toed, two two-toed) were found
between the first and second search. In no case was there evidence that two
or more sloths had defecated in exactly the same place, as noted for semiwild
three-toed sloths by Krieg (1939). Using the average defecation rate, each
would have defecated about four times during the 35-day interval. Our estimate
of the number of three-toed sloths on the area was 12 fecal masses per 4.4 defeca-
tions = 2.4 three-toed sloths. Conversion to unit area gives 2.4 sloths per 0.324
ha = 8.5 three-toed sloths per hectare.
If two-toed sloths defecate in the wild at about the same rate as they do in
captivity (GofIart, 1971; Enders, 1940), each should have defecated once
every six days, or 5.8 times during the census period. Our estimate of the num-
ber of two-toed sloths on the area was 2 fecal masses per 5.8 defecations = 0.3
two-toed sloths. Conversion to unit area gives 0.3 sloths per 0.324 ha = 1.1
two-toed sloths per hectare.
The size of the area used for the census was small relative to the home range,
and the number of fecal masses found was smaller than would be desired for
an adequate census. Because we censused only one area by this method, there is
no satisfactory way of deriving an estimate of the variances associated with our
population estimates. Until better information is available, we use the above
results as best estimates of sloth population densities. However, the numbers of
radio-marked animals present in areas of known size on particular dates support
the census results. For example, a contiguous area of forest measuring 1.3 ha
contained three two-toed and ten three-toed sloths on 5 April 1971. These ani-
mals represented known population densities of 7.7 three-toed and 2.7 two-toed
sloths per hectare.
 Sloth Impact on Energy and Nutrients 89

No previous attempt to census sloths has been reported. Beebe (1926) felt
that there were 20 three-toed sloths in a square mile (0.0 13/ha), although he
saw only three. Walsh and Gannon (1967) reported rescue of 2,104 three-toed
and 840 two-toed sloths during flooding of moist tropical forest in Surinam.
An analysis of numbers of all animals rescued indicates that three-toed sloths
contributed about 15 percent of total mammalian biomass, being the commonest
species rescued (Eisenberg and Thorington, 1973); two-toed sloths contributed
about half this amount of biomass and were the fourth commonest animal res-
cued. Goodwin (1946) states that two-toed sloths are the more abundant of the
two species in Costa Rica. Enders (1935) was able to purchase twice as many
three-toed as two-toed sloths when he offered to buy live animals for the same
price each. Lundy (1952) felt that three-toed were five times as abundant as
two-toed sloths in the Canal Zone.

Discussion

Discussion will center on some features of the role of sloths in forest energy
processes. Because this synthesis is based on order of magnitude estimates and
limited data, the reader should keep in mind that sources of variance in the
estimates were ignored for the purposes of discussion.
When each three-toed sloth crops leaves at a rate of 5.1 g (dry weight)
per kg of sloth per day, the annual cropping rate is 1.9 kg of leaves of leaves
per kg of sloth. If we assume that two-toed sloths have the same rate of food
intake as three-toed sloths, but that only two-third of their diet is leaves and
the remainder fruit, then they annually crop 1.2 kg of leaves and 0.6 kg of fruit
per kg of sloth. Juveniles and young of both species were assumed to crop at half
the rate of adults.

Sloth Biomass Ratios and Total Biomass

Biomass ratios among the individuals of each sloth population were used
for estimating sloth cropping rates. Three-toed sloths were captured (Table 7-5)
in proportions of 33 adult males, 35 adult females, 12 juvenile, and 21 young
per 100 three-toed sloths. Two-toed sloths, which could not be sexed, were
captured in proportions of 94 adults and 6 juveniles per 100 two-toed sloths;
no young were captured. Total sloth biomass per unit area was estimated on the
basis of the proportion of sloths in each population class, the weight of the
average animal in each class, and the estimated numbers of sloths per hectare.
Biomass estimates per hectare of the three-toed sloth population classes were
adult female, 9.3 kg; juvenile, 1.0 kg; young, 1.3 kg. Three-toed sloths had a
total estimated biomass of 19.4 kg/ha. For two-toed sloths, biomass estimates
per hectare were 3.5 kg for adults and 0.1 kg for juveniles. The total for the
species combined was 22.9 kg/ha.
90 G. G. MONTGOMERY AND M. E. SUNQUIST

Table 7-5. Estimated proportion of the total annual leaf production cropped by
three-toed sloths from their modal trees. 1I

Estimated
Estimated Annual Leaf
Modal Tree Cropping by Production Percent
Animal Number Number Sloth (kg) (kg) Cropped
~-~~- -~-------~ ------- - -----~-------- --~-~--

202 71 0.498 25.8 1.9

205 138 0.903 51.4 1.8
206 18 1.172 17.1 6.8
208 103 1.128 56.9 2.0
219 490 0.794 64.8 1.2
221 84 0.558 58.6 0.9
225 379 0.859 11.1 7.7
227 358 0.712 112.2 0.6
231 205 0.586 78.0 0.7

" Annual cropping by each sloth was estimated on the basis of the frequency with which
the sloth was radio located in its modal tree. Leaf production by modal trees was estimated
on the basis of the cross-sectional area of the crown, and the proprtion of total cross-sectional
area per hectare which it represented.

The nonvolant mammalian biomass of BCI has been estimated as 53 kg/ha

(Eisenberg and Thorington, 1973), with 60 percent contributed by arboreal
species. Sloths thus contribute about 73 percent of the arboreal mammalian
biomass, while Howler monkeys (Alouatta palliata) contribute only about 12
percent (Eisenberg and Thorington, 1973). By comparison with arboreal pri-
mate herbivores elsewhere, Colobus forms 79 percent of arboreal mammalian
biomass in Ghanian forest, while Presbytis forms 92 percent of that on Ceylon
(Eisenberg et ai., 1972).
Sloth densities are roughly equivalent to those of single species of terrestrial
grazing mammals, in terms of biomass, or are higher. Mentis (1970) calculated
species biomass for data on the numbers of several species of East African large
mammals removed during a control removal operation. The commonest animal
killed, the zebra (Equus), was apparently present in densities equivalent to 23
kg/ha. White rhino (Ceratotherium simum) were present at about 25 kg/ha,
while all other species biomasses of grazers and browsers were less than 10
kg/ha.
Total terrestrial herbivore biomass on two Ceylonese national parks was
8.6 kg/ha and 7.5/ha (McKay and Eisenberg, 1973). Elephants (Elaphas
maximus) contributed the greatest species biomass (4.0 kg/ha) in one area
(McKay, 1973), while axis deer (Axis axis) contributed the greatest (2.6 kg/hal
in the other (Eisenberg and Lockhart, 1972).
The total animal biomass in Puerto Rican forest has been estimated at 68
kg/ha (Odum et ai., 1970) and the arboreal herbivores as 3.21 kg/ha. Sloths
on BCI represent a biomass about six times as great as the arboreal animal
biomass (including rats) in Puerto Rican forest. Finally, the BCI sloth biomass
 Sloth Impact on Energy and Nutrients 91

is nearly equal to the total herbivore biomass of 30 kg/ha given for a central
Amazonian forest ecosystem (Fittkau and Klinge, 1973).

Total Cropping Rate by Sloths

Cropping rate by sloths per heetare of forest was estimated by multiplying

the biomass for each population class by the annual cropping rate and summing
the results. Three-toed sloth adults cropped 31.5 kg and juveniles and young
2.2 kg, giving a total cropping rate of 33.6 kg/ha/yr. Adult two-toed sloths
cropped 4.3 kg and juveniles 0.04 kg, for a total rate of 4.3 kg/ha/yr. Sloths in
aggregate cropped leaves at an annual rate of 38 kg/ha.
Howler monkeys, despite lower density, lower biomass. and lower propor-
tion of leaves in the diet, may annually crop about one-third more leaves than
sloths. Based on the figures given above for a leaf intake rate of 38 g/kg of
howler monkey per day, or 13.9 kg of leaf per kg of howler monkey per year,
and on the estimated biomass of 3.8 kg of howler monkey per hectare, they may
crop 52.8 kg of leaves per hectare of forest per year.
This estimated leaf cropping rate is based on combining several field esti-
mates which were themselves subject to large sources of error, and which need
validation. The estimated biomass and particularly the estimated leaf intake
rates may be higher than they in fact are for the BCI howler monkey popula-
tion. If this is not the case, then a higher total leaf cropping rate by howler
monkeys may be related to (1) higher metabolic demands imposed by being
homiotherms, and (2) greater amount of leaf required by an animal with a gut
less specialized for ruminant-like digestion (Hladik and Hladik, 1969) to derive
the same nutritional benefits as sloths, which are relatively specialized for such
digestion (Denis et aI., 1967; Goffart, 1971).

Proportion of Forest Leaf Production Cropped by Slotbs

Sloths cropped an estimated 0.63 percent of the total annual leaf production
of the forest (Figure 7-6). By comparison, Odum and Ruiz-Reyes 1970) esti-
mated total cropping rates by folivorous insects at about 7 percent. For temperate
forest Bray (1964) reported mean annual utilization of portions of attached
leaves by insects ranging from 5.9 to 10.6 percent of the total leaf surface.
Hladik and Hladik (1972, Table 3) estimated amounts of leaf and other food
eaten by Presby tis senex on Ceylon and the total production of leaves by various
food species. Based on those data, the monkeys cropped from 5.5 to 29.5 per-
cent of the total leaf production on their territories. Furthermore, the data would
indicate that from 15 to 47 percent of the annual leaf production of all Adina
cordi/alia (Rubiaceae) on each territory was ingested by the Presby tis sellex
that lived on that territory, as well as high percentages of the annual leaf pro-
duction of other tree species. These cropping rates seem high at least by order
92 G. G. MONTGOMERY AND M. E. SUNQUIST

Sloths: -38 kg of Leaves/ha/yr (0.6%)

1.6% (38 kg/ha/yr) 0.8% (19 kg/ha/yr)

-:::::======:::::;;:::==:::) Total Sloth Biomass
(22.88 kg/hal and
Maintenance

0.8% (19 kg/ha/yr)

Returned as Feces

Decomposes Very Slowly

Fig. 7-6. Schematic summary of the proportion of the total forest leaf production
that sloths annually crop (top) and the proportion of the leaf production of those
trees used by sloths that is cropped annually by both sloth species in aggregate (bot-
tom). Both the proportions of leaf production cropped (shown as percentages) and
the amount per unit area (kilogram per hectare) are shown. Note that half of the
leaf cropped by sloths is returned to the forest as slowly decomposing feces.

of magnitude, particularly when it is considered that other animals, vertebrate

and invertebrate, probably cropped leaves from those trees. Validation of these
results by more exact methods is necessary before it can be concluded that a
single species of mammalian arboreal foliovore crops a percentage of its food
supply comparable to percentages cropped by terrestrial foliovores.
Terrestrial grazers in aggregate crop as much as 45 percent of the annual
net production on managed rangeland (reviewed by Wiegert and Evans, 1967)
and 30 percent or more from wild grassland. A single wild species of large mam-
malian terrestrial grazer, such as the kob (Adenota kob), may ingest 10 per-
cent of the primary production (Buechner and Golley, 1967). However, two
to five terrestrial species usually total at least 70 percent of the herbivore biomass,
even though there may be as many as 18 herbivorous species present (McKay
and Eisenberg, 1973). High total cropping values for terrestrial mammalian
grazers in the tropics are primarily due to several species cropping a particular
area, rather than to high cropping rates by a single species.
 Sloth Impact on Energy and Nutrients 93

Cropping of Trees Used by Sloths

Many quantitative and qualitative factors affect the suitability of plants as

food for a sloth or other herbivore (McCullough, 1970). Depending on nutri-
tional requirements and digestive capabilities, each species will select more or
less different foods, or will select them in different proportions, from a given
area in which several species of herbivore coexist (cf. Hladik et al., 1971). Both
the kind and amount of food the herbivore selects may be an important facet of
the life history strategy (cf. McKay, 1973, p. 95).
This selection of food makes it somewhat unrealistic to compute cropping
rates on the basis of total forest leaf production. More meaningful figures for the
impact of the sloths on the forest can perhaps be derived by relating the crop-
ping rate to the portion of the forest used by sloths. Sloths used 25 percent of
the trees on our sample areas; these trees produce 40 percent of the total annual
leaf production, or 2,400 kg of leaves per hectare. In cropping these trees,
sloths would remove an estimated 1.6 percent of the annual leaf production
(Figure 7-6).
Even more intensive cropping of a tree results when it is the modal tree of
a sloth. We estimated the total annual production of each of the modal trees
and the proportion of the year the sloth spent in the tree on the basis of radio
locations Table (7-5). Some modal trees lost as much as 7.7 percent of their
estimated annual leaf production to a single sloth (Fig. 7-7). In addition, some
modal trees were used by other sloths, increasing the total cropping of the tree.
The maximum estimated total cropping for a single tree was 20 percent annually
for a small Lacmellea panamensis, which received 87 sloth-days of use by a total
of four animals.

?- _______-,<: 7.7% «
~
1.2 kg/yr)
)
<3.8% «0.6 kg/yr)
Individual Sloth
Growth and
Maintenance

Modal Tree

<3.8% «0.6 kg/yr)

Returned as Feces

Decomposes Very Slowly

Fig. 7-7. Schematic summary of the proportion of the annual leaf production of
individual trees that are used most often by a sloth (modal tree) that is cropped by
the sloth. Note that half the leaf material is returned to the modal tree as slowly
decomposing feces.
94 G. G. MONTGOMERY AND M. E. SUNQUIST

Role of Sloths in Forest Processes

Intensive cropping of a tree by sloths is not entirely disadvantageous because

the sloth selectively returns waste material to it. We estimate that half the leaf
material eaten by sloths is returned to the forest system as feces. Fecal masses
are selectively placed near the boles of trees most often used by sloths. Each
fecal mass is a package of nutrients and minerals that, as it leaches and decom-
poses, becomes available to the plants near it. The uric acid concentration of
sloth urine is relatively high (Goffart, 1971), and much of the nitrogenous end
products of digestion are passed in urine. Presumably these products are recycled
relatively quickly from the fecal mass. Other digestive by-products and undigested
leaf passed as feces would become available more slowly.
Roots of tropical trees may selectively grow toward sources of nutrients
(Odum, 1970b) and form an absorptive mat at the nutrient source (Odum,
1970c), possibly as an adaptation where there are mineral deficiencies. Trees that
have sloth feces near their boles may have an advantage in competing for nu-
trients and minerals from sloth feces because their roots have shorter distances
to grow than the roots of neighboring trees.
Of themselves, the amounts of nutrients and minerals that reach a tree from
sloth feces may not appear to be of sufficient quantity to affect the tree. How-
ever, sloth feces represent one of the few stable, long-term sources of nutrients
and minerals in the forest system. Partly because three-toed sloths bury their
fecal masses, sloth feces remain where they are placed by the animal for long
periods of time (six or more months). The feces decomposes very slowly in
place; we could readly recognize more than 90 percent of fecal masses more than
six months after they were originally found, and much of the original weight
of the feces remained. By comparison, leaves decompose rapidly in moist tropical
forest (Jenny et aI., 1949). For example, Luse (1970) noted that leaves of
Sloanea and Dacryodes were completely decomposed 51 days after being placed
on the forest floor, and noted that minerals (P) moved laterally (downstream)
from where the leaves decomposed.
In a system where most of the nutrients and minerals are bound up in plant
and animal tissue and the remainder recycled rapidly, sloth feces stands out as
a long-term stable source that is available only to certain trees. These trees pay
the price of being cropped by the sloths but receive a limited resource in return.
The effect of sloth feces may be related to stabilizing some components of the
forest system. Jordan et al. (1972) indicate that low rates of flow of minerals
relative to input rates will result in more stable systems; sloths slow the normally
high recycling rates for certain trees. Sloths may thus represent not a chemical
but a behavioral specialist that (Odum, 1970, p. 1-242): " ... provide the forest
special services in holding or recycling particular elements."
In the coevolution of sloths and their food trees, the rate at which trees
evolve defense mechanisms against sloths may have been slowed by two factors.
In reducing the cropping pressure on single species by utilizing a broad spectrum
(each three-toed sloth tends to use a different species as a modal tree), predator
pressure against a single tree species is reduced. By selectively returning feces
 Sloth Impact on Energy and Nutrients 95

to the cropped trees sloths may promote growth and survival of the trees, even
while they act as predators on them. This situation differs from the influence
of mammalian seed predators (Smith, 1970) where predation may relatively
rapidly influence the pattern of reproduction of the trees.

Acknowledgment

The study was supported by Smithsonian Research Foundation Grant No.

435090 to J. F. Eisenberg, and by the Environmental Sciences Program, Smith-
sonian Institution. Smithsonian Tropical Research Institute provided logistic
support, and personnel of the Institute gave advice and field help. R. A. Foster
identified trees used by study sloths and gave other invaluable field help and
advice. Judith A. Montgomery prepared the histological material for identifica-
tion of stomach contents and assisted with fieldwork. Many visitors to Barro
Colorado Island contributed by their comments while work was in progress,
and in some cases helped to capture sloths. E. Leigh offered welcome comments
on the manuscripts.

References

Baumgartner, L. L., and A. C. Martin. 1939. Plant histology as an aid in squirrel

food-habits studies. J. Wildlife Mgmt. 3 :266-268.
Beebe, W. 1926. The three-toed sloth, Bradypus cuculliger cuculliger Wagler. Zo-
ologica 7: 1-67.
Bennett, C. F., Jr. 1963. A phytophysiognomic reconnaissance of Barro Colorado
Island, Canal Zone. Smith. Misc. Coli. 145, 7: 1-8.
Bourliere, F. 1973. The comparative ecology of rain forest mammals in Africa and
tropical America: Some introductory remarks. In Tropical Forest Ecosystems
in Africa and South America: A Comparative Review, B. J. Meggers, E. S.
Ayensu, and W. D. Duckworth, eds. pp. 279-292. Washington, D.C.: Smith-
sonian Inst. Press.
Bray, J. R. 1964. Primary consumption in three forest canopies. Ecology 45: 165-
167.
Buechner, H. K., and F. B. GoHey. 1967. Preliminary estimation of energy flow in
Uganda kob (Adenota kob thomasi Newman). In Secondary Productivity of
Terrestrial Ecosystems, K. Petrusewicz, ed. pp. 243-260. Vol. 1, pp. 1-379.
Warsaw: Inst. Ecology Polish Acad. Sci.
Carpenter, C. R. 1934 (1964). A field study of the behavior and social relations of
howling monkeys. (Alouatta palliata). In Naturalistic Behavior of Nonhuman
Primates, C. R. Carpenter, ed. pp. 3-92. University Park, Pa.: Pennsylvania
State Univ. Press.
Carvalho, C. T. de. 1960. Notes on the three-toed sloth, Bradypus tridactylus. Mam-
malia 24: 155-156.
96 G. G. MONTGOMERY AND M. E. SUNQUIST

Cloudsley-Thompson, J. L. 1972. The habitat and its influence on the evolutionary

development. In Biology of Nutrition, R. N. T.-W. Fiennes, ed. pp. 351-373.
New York: Pergamon Press.
Denis, c., C. Jeuniaux, M. A. Gerebtzoff, and M. Goffart. 1967. La digestion stoma-
cale chez un paresseux: l'unau Choloepus hofJmanni Peters. Ann. Soc. ZooT.
Belg.97:9-29.
Eisenberg, J. F., and M. Lockhart. 1972. An ecological reconnaissance of Wilpattu
National Park, Ceylon. Smithsonian Contr. Zoo!. 101.
- - - , and G. M. McKay. 1974. Comparison of ungulate adaptations in the New
World and Old World tropical forests with special reference to Ceylon and the
rainforests of Central America. In The Behavior of Ungulates and Its Relation
to Management, V. Geist and F. Walther, eds. Int. Union Conservation Nature
(in press).
- - - , N. A Muckenhim, and R. Rudran. 1972. The relation between ecology and
social structure in primates. Science 176: 863-874.
- - - , and R. W. Thorington. 1973. A preliminary analysis of a neotropical mam-
mal fauna. Biotropica 5: 150-161.
Enders, R. K. 1935. Mammalian life histories from Barro Colorado Island, Panama.
Bull. Mus. Compo Zool. 78:385-502.
- - - . 1940. Observations on sloths in captivity at higher altitudes in the tropics
and in Pennsylvania. J. Mammal. 17:165-166.
Fittkau, E. J., and H. Klinge. 1973. On biomass and trophic structure of the central
Amazonian rain forest ecosystem. Biotropica 5: 1-14.
Goffart, M. 1971. Function and Form in the Sloth. 225 pp. New York: Pergamon
Press.
Golley, F. B. 1967. Methods of measuring secondary productivity in terrestrial ver-
tebrate populations. In Secondary Productivity of Terrestrial Ecosystems, K.
Petrusewicz, ed. pp. 99-124. Inst. Ecol. Polish Acad. Sci.
- - - . 1972. Summary. In Tropical Ecology with an Emphasis on Organic Produc-
tion, P. M. Golley and F. B. Golley, compilers. pp. 407-413. Athens: Inst. Ecol.,
Univ. Georgia.
Goodwin, G. G. 1946. Mammals of Costa Rica. Bull. Am. Mus. Nat. Hist. 87:275-
478.
Harrison, J. L. 1972. Feeding habits of animals in arboreal habitats. In Biology of
Nutrition, R. N. T.-W. Fiennes, ed. pp. 505-512. New York: Pergamon Press.
Hladik, A, and C. M. Hladik. 1969. Rapports trophiques entre vegetation et primates
dan la foret de Barro Colorado (Panama). Terre Vie 1:29-117.
Hladik, C. M., J. Bousset, P. Valdebouze, G. Viroben, and J. Delort-Laval. 1971.
Le regime alimentaire des primates de lile de Barro-Colorado (Panama). Folia
Primat. 16:85-122.
- - - , and A Hladik. 1972. Disponibilites alimentaires et domaines vitaux des pri-
mates a Ceylan. Terre Vie 2: 149-215.
Hofman, R. R., and D. R. M. Stewart. 1972. Grazer or browser; a classification based
on the stomach-structure and feeding habits of East African ruminants. Mam-
malia 36: 226-240.
Hussain, M., B. N. Majumdar, B. Sahai, and N. D. Kehar. 1960a. Studies on tree
leaves as cattle fodder. II. The nutritive value of Bargad leaves (Ficus ben-
galensis).Ind. J. Dairy Sci. 13: 1-8.
- - - . 1960b. Studies on tree leaves as cattle fodder. IV. The nutritive value of Pipal
leaves (Ficus religiosa). Ind. J. Dairy Sci. 13:9-15.
 Sloth Impact on Energy and Nutrients 97

Jenny, H., S. P. Gessel, and F. T. Bingham. 1949. Comparative study of decomposi-

tion rates of organic matter in temperate and tropical regions. Soil Sci. 68 :419-
432.
Jordan, C. F., 1. R. Kline, and D. S. Sasscer. 1972. Relative stability of mineral
cycles in forest ecosystems. Amer. Natur. 106:237-253.
Kleiber. M. 1961. The Fire of Life. 454 pp. New York: John Wiley and Sons.
Krieg, H. 1939. Begegnungen mit Ameisenbaren und Faultieren in freier Wildbahn.
Z. Tierpsychol. 2: 282-292. .
- - - . 1961. Das Verhalten der Faultiere (Bradypodidae). Hanbllch Zool. 8/27.
10(12b) :20-23.
Lieth, H. 1970. The water content of some leaves at El Verde. Appendix A. In
A Tropical Rain Forest. H. T. Odum, ed. p. 1-281. Washington, D.C.: Div.
Tech. Information. U.S. Atomic Energy Comm.
Lundy. W. E. 1952. The upside-down animal. Nat. Hist. 61: 114-119.
Luse, R. A. 1970. The phosphorus cycle in a tropical rain forest. In A Tropical Rain
Forest, H. T. Odum, ed. pp. H-161 to H-166. Washington, D.C.: Div. Tech. In-
formation, U.S. Atomic Energy Comm.
McCrane, M. P. 1966. Birth, behaviour and development of a hand-reared two-toed
sloth, Choloeplls didactyills. Intern. Zoo Yearbook 6: 153-163.
McCullough, D. R. 1970. Secondary production of hirds and mammals. In Analysis
of Temperate Forest Ecosystems, D. E. RichIe. ed. pp. 107-130. New York:
Springer-Verlag.
McKay. G. M. 1973. Behavior and ecology of the Asiatic elephant in southeastern
Ceylon. Smithsonian ContI'. Zool. 125.
- - - , and 1. F. Eisenberg. 1974. Movement patterns and habitat utilization of un-
gulates in Ceylon. In The Behavior of Ungulates and Its Relation to Manage-
ment, V. Geist and F. Walther. eds. LU.C.N. (Pub!.) (in press).
Mentis, M. T. 1970. Estimates of natural biomasses of large herbivores in the Um-
folozi Game Reserve area. Mammalia 34:363-393.
Montgomery, G. G. 1969. Weaning of captive raccoons. J Wildlife Mgmt. 33:154-
169.
- - - , W. W. Cochran, and M. E. Sunquist. 1973. Radio-locating arboreal verte-
brates in tropical forest. J. Wildlife Mgmt. 37:426-428.
Odum, H. T. 1970a. Summary: An emerging view of the ecological system at El
Verde. In A Tropical Rain Forest, H. T. Odum, ed. pp. 1-191 to 1-289. Wash-
ington. D.C.: Div. Tech. Information, U.S. Atomic Energy Comm.
- - - . 1970b. Introduction to section F. In A Tropical Rain Forest, H. T. Odum,
ed. pp. F-3 to F-7. Washington, D.C.: Div. Tech. Information, U.S. Atomic
Energy Comm.
- - - . 1970c. Rain forest structure and mineral-cycling homeostatis. In A Tropi-
cal Rain Forest. H. T. Odum, ed. pp. H-3 to H-52. Washington, D.C.: Div.
Tech. Information, U.S. Atomic Energy Comm.
- - - , W. Abbott, R. K. Selander, F. B. Golley. and R. F. Wilson. 1970. Estimates
of chlorophyll and biomass of the Tabonuco forest of Puerto Rico. In A Tropi-
cal Rain Forest, H. T. Odum, ed. pp. 1-3 to 1-19. Washington, D.C.: Div. In-
formation, U.S. Atomic Energy Comm.
- - - , and 1. Ruiz-Reyes. 1970. Holes in leaves and the grazing control mechanism.
In A Tropical Rain Forest, H. T. Odum, ed. pp. 1-69 to 1-80. Washington, D.C.:
Div. Tech. Information, U.S. Atomic Energy Comm.
Overton, W. S., and D. E. Davis. 1969. Estimating the numbers of animals in wildlife
98 G. G. MONTGOMERY AND M. E. SUNQUIST

populations. In Wildlife Management Techniques, R. H. Giles, Jr., ed. pp. 403-

456. Washington, D.C.: The Wildlife Society.
Ovington, J. D., and J. S. Olson. 1970. Biomass and chemical content of EI Verde
lower montane rain forest plants. In A Tropical Rain Forest, H. T. Odum, ed.
pp. H-53 to H-61. Washington, D.C.: Div. Tech. Information, U.S. Atomic
Energy Comm.
Ozorio de Almeida, A., and Branco de A. Fialho. 1924. Metabolisme, temperature et
quelques autres determinations physiologiques faites sur Ie paresseux (Bradypus
tridactylus). Compo Rend. Soc. Bioi. 91:1124-1125.
Petersen, J. C. B., and R. L. Casebeer. 1971. A bibliography relating to the ecology
and energetics of East African large mammals. E. Afr. Wildlife J. 9: 1-23.
Smith, C. C. 1970. The coevolution of pine squirrels (Tamiasciurus) and conifers.
Ecol. Monographs 40:349-371.
Standley, P. C. 1933. The flora of Barro Colorado Island, Panama. Contr. Arnold
Arboretum 5: 1-178.
Stewart, D. R. M. 1967. Analysis of plant epidermis in faeces: A technique for study-
ing the food preferences of grazing herbivores. 1. Appl. Ecol. 4: 83-111.
Sunquist, M. E., and G. G. Montgomery. 1973. Activity patterns and rates of move-
ment of two-toed and three-toed sloths (Choloepus hojjmanni and Bradypus
infuscatus). J. Mammal. 54:946-954.
Tadaki, Y. 1966. Some discussions on the leaf biomass of forest stands and trees.
Bull. Gov. For. Exp. Sta. (Tokyo) 184: 135-161.
Van Doorn, C. 1971. Verzogerter Zyklus der Fortpflanzung bei Faultieren. Z. Kainer
Zoo 14:15-22.
Veselovsky, Z. 1966. A contribution to the knowledge of the reproduction and growth
of the two-toed sloth, Choloepus didactylus, at the Prague Zoo. Intern. Zoo
Yearbook 6:147-153.
Walsh, J., and R. Gannon. 1967. The Time Is Short and the Water Rises. 224 pp.
London: Thomas Nelson and Sons.
Wiegert, R. G., and F. C. Evans. 1967. Investigations of secondary productivity in
grasslands. In Secondary Productivity of Terrestrial Ecosystems, K. Petrusewicz,
ed. pp. 499-518. Inst. Ecol. Polish Acad. Sci. V 2: 383-879.
Worman, R. G. 1946. Swimming Sloth. Nat. Hist. N.Y. 55:49.