Coral Disease Outbreak: Pattern, Prevalence and Transmission in
Coral Disease Outbreak: Pattern, Prevalence and Transmission in
ABSTRACT: A rapidly progressing disease outbreak affecting Acropora cervicornis (staghorn coral)
was recently observed over a wide geographic range (> 200 km) in the Florida Keys, USA. Original
observations made at White Bank Dry Rocks revealed colonies with tissue rapidly sloughing from
multifocal lesions. Over the course of the epizootic at least 72% of tagged colonies were affected, live
tissue cover of these colonies decreased from 96 ± 1% to 12 ± 5% (mean ± SE), and 28% of the
colonies suffered complete mortality. Observed rates of tissue loss were highly variable, ranging from
2 to 43 cm2 d–1, but the mean was a rapid 13 ± 11 cm2 d–1 or about 4 cm branch length d–1. Field exper-
iments demonstrate that this disease is transmissible not only by direct contact between affected and
healthy staghorn coral tissue but also by the corallivorous snail Coralliophila abbreviata. The disease
was also transmissible, although less effectively so, to the congener A. palmata (elkhorn coral). No
transmission was observed in indirect contact treatments designed to simulate diver interaction.
Transmissibility implies that the condition is indeed a biotic disease and the demonstration of effec-
tive vector transmission suggests that predation may exacerbate disease outbreaks in remnant
Caribbean acroporid populations, further impeding their potential recovery.
to 3 m, has a relatively sparse population of small Acro- ease exposed) at the end of each 3 wk trial were
pora cervicornis colonies interspersed with small head analyzed using Fisher’s exact 1-tailed test (α = 0.05;
corals (Diploria spp, Porites spp. and Montastraea StatSoft 2001).
annularis) and gorgonians. In June 2002, 19 colonies of In replicates that responded, response time was cal-
A. cervicornis were haphazardly chosen and tagged culated from the day of setup (Day 0) to the day that
for periodic monitoring. Colony condition was assessed tissue sloughing was first observed. For the predator
by visually estimating the percentage of each colony feeding treatments (see last subsection below), Day 0
that was covered with live tissue (hereafter ‘% live’). was considered as the day the feeding scar was first
Colonies were re-surveyed on November 21, 2002 and observed and was at least 2 d (range 1 to 3 d) after the
April 17, 2003, when rapid tissue loss was observed. snail had been placed on the fragment.
Thereafter, tagged colonies were surveyed more fre- Area of tissue loss was estimated from scaled photo-
quently (April 29, 2003 and May 13, 2003) to track the graphs of 17 time intervals of 10 fragments with dis-
short-term trends of this outbreak, and then were not ease response in Expts I and IV (see following subsec-
surveyed again until February 4, 2004. Prevalence of tions). Image analysis software (Image Tool Version
disease (i.e. proportion of tagged colonies with rapid 3.0, UTHSCSA) was used to measure the length and
tissue loss) was determined at each census. Transect diameter of the areas with tissue loss, and a surface
surveys conducted in the vicinity of the tagged area was calculated based on the formula for an open-
colonies confirmed similar trends among the surround- ended cylinder (πdh). The rate of tissue loss (cm2 d–1)
ing population between summer 2002 and April 2003 was calculated using the number of days between
(authors’ unpubl. data). photographs.
Qualitative surveys of Acropora cervicornis (and, at Expt I. Acropora cervicornis direct contact: Appar-
1 site, A. prolifera, an F1 hybrid of A. palmata and A. ently healthy A. cervicornis fragments were attached
cervicornis: Vollmer & Palumbi 2002) populations were in direct contact with healthy and diseased A. cervi-
made between May 2003 and August 2003 at 17 sites cornis colonies. Fragments were affixed using cable
over a wide geographic range from Biscayne National ties to branches with actively sloughing tissue (‘treat-
Park to Dry Tortugas National Park including the ment’) or branches of apparently entirely healthy
FKNMS (Table 1). Sites with known A. cervicornis (‘control’), standing A. cervicornis host colonies. Of the
occurrence were targeted because of its extremely 6 treatment fragments in the first trial, 3 were reposi-
patchy and sparse distribution. Surveys were based on tioned after 10 d because the disease front on the host
a single visit when the prevalence of rapid tissue loss colony had either lost contact with the fragment or
was ranked as high (> 40% of colonies affected), appeared inactive. Fragments in the second trial were
moderate (10 to 40%) or low (<10%) for each site. For more accurately positioned on the host branches using
sites where signs of rapid tissue loss were observed but additional cable ties to effect more consistent contact
fewer than 10 colonies were found, prevalence was with the diseased branch, so repositioning was not
conservatively ranked as low. necessary. Because the number of colonies with active
Transmission experiments. We conducted 4 sepa- disease in the surrounding population had diminished
rate controlled experiments at White Bank, in the by the second trial, it was necessary to place more than
vicinity of the tagged colonies at 2 to 3 m depth, to 1 fragment on a single standing host colony, but repli-
determine if the condition was transmissible. Loose, cates were always on separate branches.
apparently healthy, fragments of Acropora cervicornis Expt II. Acropora palmata direct contact: No A.
found at White Bank were collected and allocated to 1 palmata fragments were available at White Bank, so
of 3 experiments. For an additional experiment, frag- apparently healthy fragments were collected from Sand
ments of A. palmata were imported from other sites Island reef (25° 01.107 N, 80° 22.044 W) for the first trial
(see Expt II below). Latex gloves were worn and and Little Grecian reef (25° 07.220 N, 80° 18.050 W) for
changed between handling of replicates to minimize the second trial. These fragments were attached to dis-
cross-contamination. eased (‘treatment’) or healthy (‘control’) A. cervicornis
Each experiment was run in 2 trials, with the second colonies at White Bank using cable ties. In some cases,
trial undertaken to improve treatment exposure in the the size or shape of the A. palmata fragment made
direct and indirect contact experiments (see following it impossible to attach to a standing A. cervicornis
subsections) and to increase replication (see Table 2 colony. In these cases, the A. palmata fragments were
for sample sizes) in the predator experiment. In each stabilized on the substrate using cable ties and a treat-
trial, the replicates were checked for response (defined ment (diseased) or control (apparently healthy) A. cer-
as the observation of tissue sloughing) every 2 to 6 d vicornis fragment was cable tied to make direct contact
for a period of 3 wk. Frequencies of positive (disease) between the fragments. As with Expt I, fragments were
responses among treatments (healthy control and dis- more carefully positioned on control and treatment
122 Mar Ecol Prog Ser 301: 119–128, 2005
RESULTS
Transmission experiments
Field surveys
No control replicates showed disease response in any
The fate of the tagged colonies at White Bank is of the 4 experiments. The first trial of each direct con-
illustrated in Fig. 2. During the baseline period (June tact experiment (Expts I and II) produced a response
to November 2002), a few colonies showed typical but it was not statistically significant (Table 2). Our
signs of predation by Hermodice carunculata (fire- observations suggested that non-responding replicates
worm) and Coralliophila abbreviata; 1 colony was often had not maintained consistent contact with the
nearly killed by unknown causes, and 1 colony dis- actively diseased area. In the second trials of Expts I
played rapid tissue loss. During April to May 2003, and II, more replicates responded and response was
an average of 1 new colony (5% of tagged) wk–1 faster, suggesting that direct tissue contact was impor-
became affected by rapid tissue loss and 3 colonies tant in transmission. The second trial of Expt I yielded
died completely. By the final census (February 2004), significant response (p = 0.0455). In Expt II, transmis-
a total of 5 colonies were completely dead and, of the sion did occur in 3 of 5 second trial treatment replicates,
13 remaining colonies with live tissue (1 was lost), 2 although it was not statistically significant (p = 0.2857).
still showed active signs of rapid tissue loss. During Once rapid tissue loss was transmitted to Acropora
the course of the outbreak, a total of 13 colonies had palmata (Expt II), signs were identical in appearance to
been directly observed to be affected by rapid tissue those in A. cervicornis, namely irregular tissue slough-
loss. Of the 5 colonies that were not directly ing with no resemblance to a band (Fig. 4).
observed with disease signs, 2 were dead and the Indirect contact (Expt III) failed to produce disease
other 3 had less than 10% live tissue cover, suggest- signs in any of the treatment or control replicates, even
ing that they had indeed been affected by rapid in the second trial where tissue was clearly injured by
tissue loss at some point. the swabbing. In contrast, feeding by treatment snails
Williams & Miller: Disease outbreak in Acropora cervicornis 123
DISCUSSION
Experimental response
Fig. 4. Acropora palmata. Healthy fragment in direct contact with (a) healthy and (b) diseased A. cervicornis fragments in Expt II.
At initial setup (Day 0) both treatments were healthy; 4 d later the A. palmata fragment in the disease-exposed treatment is almost
completely dead. Note the sloughing tissue on A. palmata does not form ‘band-like’ appearance typical of white band disease
gloves which was exposed to the water column in the In 1977, Acropora cervicornis populations in the Dry
indirect inoculation (Expt III), snails may have effec- Tortugas rapidly sloughed live tissue following an
tively retained the transmissive agent by withdrawing extreme cold front when temperatures dropped to
their guts (which are everted when feeding, Ward 14°C (Davis 1982), suggesting that the condition at
1965) during transfer from the inoculation colony to the White Bank could have resulted from physiological
treatment fragment. stress in response to anomalous water temperatures or
other environmental factors. However, colonies con-
tinued to become affected at a high rate throughout
Acute stress? the spring and summer outbreak of 2003 (Figs. 2 & 3),
a few colonies manifested active disease in winter
Tissue sloughing similar to that seen at White Bank (February 2004, Fig. 2), and many diseased A. cervi-
has been associated with acute environmental stress. cornis colonies were observed at White Bank in
Table 2. Acropora cervicornis (Ac) and Acropora palmata (Ap). Results of Expts I to IV (Trials 1 and 2); n: number of replicates;
R: number of replicates showing disease response (subscripted numbers are trial numbers); R time: range in response time, or
number of days from the initial exposure to the first sign of disease for all responding replicates; na: not applicable. Responses
were compared using Fisher’s exact test; p-values are shown for each trial. Trials 1 and 2 of Expt IV were pooled since methods
were identical
summer 2004 (D. E. Williams pers. obs.). The sustained branches…which enlarged through necrosis. [In A.
and ongoing nature of the epizootic and the persis- cervicornis]…necrosis spread rapidly over the branches
tence of active disease signs under a range of seasonal and fragments completely covered with living tissue
conditions, indicate that the rapid tissue loss of A. (length 20 cm) were commonly found wholly dead one
cervicornis observed at White Bank (and probably at week later’. Because their report roughly coincided
other sites in the Florida Keys, Fig. 3) resulted from a with the first descriptions of WBD, it was probably
biotic, transmissible disease, rather than an acute assumed to be WBD (Green & Bruckner 2000). How-
physiological stress response to environmental factors. ever, Bak & Criens’ (1981) description of disease signs
There remains the possibility that subtle environmen- on A. cervicornis is strikingly similar to that reported
tal stresses may have influenced pathogen virulence herein and to published reports of A. palmata afflicted
and/or host susceptibility in the onset of the outbreak. with white pox (Rodríguez-Martínez et al. 2001, Patter-
son et al. 2002, Sutherland et al. 2004).
LITERATURE CITED
population collapse: post-hurricane predation on Jamaican Rodríguez-Martínez RE, Banaszak AT, Jordan-Dahlgren E
staghorn corals. Smithson Contrib Mar Sci 31:25 (2001) Necrotic patches affect Acropora palmata (Sclerac-
Marsden JR (1962) A coral-eating polychaete. Nature 193:598 tinia: Acroporidae) in the Mexican Caribbean. Dis Aquat
McCarthy M (2001) A century of the US Army yellow fever Org 47:229–234
research. Lancet 357:1772 Santavy DL, Peters EC (1997) Microbial pests: coral disease in
Miller MW, Baums IB, Williams DE, Szmant A (2002) Status of the Western Atlantic. Proc 8th Int Coral Reef Symp 1:
candidate coral, Acropora palmata, and its snail predator 607–612
in the upper Florida Keys National Marine Sanctuary: Santavy DL, Peters EC, Kozlowski J, Wilkinson SS (1995)
1998–2001. NOAA Tech Memo NMFS-SEFSC-479 Characterization of the bacterium suspected in the inci-
Patterson KL, Porter JW, Ritchie KB, Polson SW, Mueller E, dence of white band disease. Abstr Gen Meet Am Soc
Peters EC, Santavy DL, Smith GW (2002) The etiology of Microbiol 95(N-1):332
white pox, a lethal disease of the Caribbean elkhorn coral, StatSoft (2001) STATISTICA (data analysis software system),
Acropora palmata. Proc Natl Acad Sci USA 99:8725–8730 Version 6, Tulsa, OK. www.statsoft.com
Peters EC, Oprandy JJ, Yevich PP (1983) Possible causal Sussman M, Loya Y, Fine M, Rosenberg E (2003) The marine
agent of ‘white band disease’ in Caribbean acroporid fireworm Hermodice carunculata is a winter reservoir and
corals. J Invertebr Pathol 41:394–396 spring–summer vector for the coral-bleaching pathogen
Precht WF, Bruckner AW, Aronson RB, Bruckner RJ (2002) Vibrio shiloi. Environ Microbiol 5:250–256
Endangered acroporid corals of the Caribbean. Coral Sutherland KP, Porter JW, Torres C (2004) Disease and immu-
Reefs 21:41–42 nity in Caribbean and Indo-Pacific zooxanthellate corals.
Reyes-Bonilla H, Calderon-Aguilera LE (1999) Population Mar Ecol Prog Ser 266:273–302
density, distribution and consumption rates of three coralli- Vollmer SV, Palumbi SR (2002) Hybridization and the evolu-
vores at Cabo Pulmo Reef, Gulf of California, Mexico. Mar tion of reef coral diversity. Science 296:2023–2025
Ecol 20:347–357 Ward J (1965) The digestive tract and its relation to feeding
Richardson LL (1998) Coral diseases: What is really known? habits in the stenoglossan prosobranch Coralliophila
Trends Ecol Evol 13:438–443 abbreviata. Can J Zool 43:447–464
Ritchie KB, Smith GW (1995) Preferential carbon utilization Williams DE, Miller MW (in press) Importance of disease and
by surface bacterial communities from water mass, normal predation to the growth and survivorship of juvenile Acro-
and white-band diseased Acropora cervicornis. Mol Mar pora palmata and Acropora cervicornis: a demographic
Biol Biotechnol 4:345–352 approach. Proc 10th Int Coral Reef Symp
Ritchie KB, Smith GW (1998) Description of Type II white Woodley C, Bruckner A, Galloway S, McLaughlin S, Downs
band disease in acroporid corals. Rev Biol Trop 46 (Suppl 5): C, Fauth J, Shott E, Larkin K (2003) Coral disease and
199–203 health: a national research plan. NOAA, Silver Spring, MD
Editorial responsibility: Charles Birkeland (Contributing Submitted: December 3, 2003; Accepted: March 3, 2005
Editor), Honolulu, Hawaii, USA Proofs received from author(s): September 13, 2005