106 
trols, were taken as the seaweed consumed by 
grazers. Weight changes in the controls were 
usually very small. 
RESULTS 
The allowance method of Tukey (1953:90) 
was used to test for differences in consumption. 
Differences between the means of algal con- 
sumption values, significant at a 5% error rate, 
are indicated by separate boxes in Figure 1. Plant 
species falling in intermediate positions and not 
significantly different from adjacent groups are 
not enclosed. In most cases three distinct levels 
of consumption are evident. The results show 
that in many cases Macrocystis was consumed in 
relatively large amounts, appearing at or near 
the highest level; Cystoseira, Gigartina, or Ptery- 
gophora were frequently, but not always, at the 
lowest level; and the other algal species were at 
intermediate levels. 
Means and ranges of quantitative data for all 
grazers are provided in Figure 2. The ranges 
are, in some instances, rather great. Variations in 
ambient temperature between 13 and 19' C 
caused fluctuation in feeding rate. Sizes of ani- 
mals, while similar, did vary. Other factors 
doubtless affected feeding rate and selection. 
Despite the variability, a- considerable degree of 
correspondence was found between experiments. 
To test the agreement of results, a ranking 
method described by Kendall (1955:95) was 
employed. In each trial, the algae were assigned 
ranks from 1 to 7 in order of decreasing quan- 
tities consumed. Values of W, the coefficient of 
concordance, ranged from 0.4 1 to 0.79 (Table 
1); all were significant beyond the .001 level. 
The order of preference of each grazing species, 
therefore, may be regarded as consistent and the 
mean values for consumption (Fig. 2) afford 
the best estimate of the relative degree of at- 
tractiveness of the algae to the grazing animals. 
The data for the entire experimental series 
are included for two species, Strongylocentrotus 
purpuratus (Table 2) and Lyte chirms anamesus 
(Table 3), in order to provide an indication of 
the intraspecific variability of results. S, pur- 
puratus collected from a variety of habitats, 
showed limited variation in preference behavior. 
Mean values for L. anamesus from the La Jolla 
PACIFIC SCIENCE, Vol. XX, January 1966 
Canyon and from the Mission Bay entrance 
show general agreement. The same tendency 
toward a specific pattern of preferential response 
to the algae tested here was found In other graz- 
ing organisms collected from a variety of plant . 
environments. ( A recent extension of these 
studies has shown similarities in response by 
Haliotis and Strongylocentrotus collected at lati- 
tudes well beyond the range of certain of the 
algae used here. Hence a species-specific pattern 
of selection appears to exist, some species being 
characteristically attractive while others are not.) 
Macrocystis was most highly preferred in 7 
of 11 cases; typical kelp bed inhabitants fall 
within this group. Egregia, a shallow water kelp 
(0-8 m) was most preferred by Norrisia , H. ful- 
gens, and Apiysia , all most common at lesser 
depths. Lytechinus , normally a relatively deep 
water dweller, preferred red algae (represented 
in these experiments by Gigartina armata). Red 
algae generally supplant brown algae In domi- 
nance at depths exceeding 35 m. 
Several species of brown algae were consist- 
ently consumed in greatest relative quantities 
by all grazers. Egregia and Laminaria, in addi- 
tion to Macro cysfisf ' were generally favored. 
Cystoseira, on the other hand, was seldom in- 
gested by grazers other than Pugettia, Taliepus, 
and H. corrngata. Gigartina was readily eaten by 
Lytechinus, Apiysia , Astraea, and Strongylocen - 
trotus, but was hardly touched by Pugettia, Talie- 
pus, Norrisia , and H. corrngata. 
The coefficient of concordance (W t ) was de- 
termined for a summation of the rankings of 
preference for all 11 grazing species (Table 4) 
and is significant at better than the 0.1% level. 
The order of preference is, therefore, generally 
consistent throughout the group of species, with 
Macrocystis as the most commonly preferred 
item and Cystoseira as the least acceptable. 
Since some closely allied species equipped 
with similar chewing, biting, or rasping organs 
exhibited divergent preference behavior, the re- 
sults are not considered to be a consequence of 
relative mechanical ability to cope with a par- 
ticular food. Toughness of foods might be sug- 
gested as influencing choice when measured by 
the techniques employed; the resilient, denser 
algae being consumed at a slower rate than the 
tenderer tissues. If this were true, one would ex- 
