|dc.description.abstract||2 The production of coral offspring and their survival through early ontogeny to sexual maturity
3 are both vitally important for the persistence of coral-dominated reefs. Understanding factors
4 which affect these processes is important where limited connectivity occurs. This is the case
5 on the high-latitude Two-mile Reef (TMR) at Sodwana Bay in South Africa. A combination of
6 in situ and ex situ experimental work was conducted, investigating factors which affect the pre7
and post-settlement stages of corals. In vitro experiments conducted on two representative
8 scleractinian broadcast spawning corals, Acropora austera and Hydnophora exesa, revealed
9 that fertilisation success in both species diminishes significantly with a reduction in sperm
10 concentration and water salinity. Fertilisation success was highest for A. austera at 106 sperm
11 ml-1 (56.46% ± 0.83, mean ± SE), and at 105 sperm ml-1 (38.76% ± 1.29) for H. exesa. At 104
12 sperm ml-1 there was a significant reduction in fertilisation of 80% and 58% for the respective
13 species. Additionally, fertilisation success of A. austera and H. exesa decreased significantly
14 by 56% and 79% respectively, when salinity was reduced by 7.06 psu. Ex situ settlement
15 experiments were then conducted to assess the settlement of their larvae in response to the
16 presence of two crustose coralline algae (CCA) a Hydrolithon sp. and Mesophyllum sp. and
17 filtered sea water (FSW) control. Settlement in the presence of Mesophyllum was not
18 significantly different from FSW, but a significant trend was observed in the presence of
19 Hydrolithon, where settlement of both corals peaked. While H. exesa post-settlement success
20 was also greatest in the Hydrolithon treatment (55.00 ± 10.47%), this was not the case for A.
21 austera, which obtained highest post-settlement success in the Mesophyllum treatment (21.67
22 ± 7.23%). However, these trends were not significant. Acropora austera and H. exesa were
23 still capable of settlement 69 and 75 days after fertilisation respectively in FSW. In general,
24 more settlement occurred on container surfaces than on CCA fragments. Results from this
25 study suggest that the inducing effect of CCA is coral and CCA taxon-specific, and that A.
26 austera planulae are more stringent in their settlement requirements.
28 In situ experiments were conducted by attaching settlement tiles to concrete Y-frames on TMR
29 for six months to assess how coral recruitment differs according to method of settlement tile
30 attachment onto concrete y-frames and how the exclusion of herbivores and predators affects
31 coral recruitment onto tiles. A new, grooved settlement tile was designed to provide refuge
microhabitats on the top surface of the tiles. In total, 32 579 recruits were detected on the
33 settlement tiles, with pocilloporids dominating the recruit composition (64%).
35 Grooved tiles were also used to assess whether coral recruit density varied between different
36 microhabitats adjacent to the tile edge (a narrow, 5 mm gap; a wide, 15 mm gap; and tiles raised
37 above the gap). Most recruitment occurred on the vertical edges and towards the edge perimeter
38 of grooved tiles regardless of treatment. The majority of recruitment on the top surface of tiles
39 occurred in the grooves (74.17%). Coral recruit densities differed significantly between the
40 three edge microhabitats, with recruit density significantly less on tiles adjacent to narrow gaps.
41 Raised tiles and tiles with a wide gap had two- and three-fold more recruits (644.33 ± 149.43
42 and 979.29 ± 170.88 recruits m-2) than tiles with a narrow gap (311.05 ± 80.82 recruits m-2).
43 This suggests that the microhabitat associated with the method of tile attachment can have a
44 significant effect on recruitment. Finally, the effect of large herbivores and predators on coral
45 recruitment and the benthic communities was assessed by placing exclusion cages over tiles.
46 Recruit densities had a two-fold reduction when herbivores and predators were excluded.
47 Additionally, CCA cover was also significantly reduced on caged tiles, and the percentage of
48 erect foliose algae, encrusting macroalgae, and turf algae was significantly greater compared
49 to uncaged tiles. This indicates that grazers may assist coral recruitment on TMR.
51 This study provided the first assessment of fertilisation success in corals at high-latitude in
52 South Africa and the results are related to information on gene flow and reef resilience. The
53 importance of suitable settlement microhabitats and grazers are also highlighted and stress the
54 need for a multi-faceted management approach to coral conservation. Furthermore, from an
55 experimental point of view, the methodological techniques used to quantify in situ recruitment,
56 such as settlement surface design and attachment technique, may have important implications
57 in quantifying recruit densities and settlement preferences. Such differences must be
58 considered when comparing the results of recruit densities in studies using dissimilar