7831

• UCREFRP Catalog Number: 7831
• Author: Dowling, T. E., et al.
• Title: Mitochondrial DNA Variability in the Endangered Razorback Sucker ‎(
• USFW Year: 1996
• Copyright Material: YES

124 mtDNA Diversity in Razorback Sucker
250 1.0
200 0.9
'" ..-,
Cl)
..... 0.8
~ <~
.~ 150 '-'
0.. >.
~ ::: 0.7
'"
..... 100 1990 ....
0 Cl)
I >
0 a 0.6
z
50
0.5
0
70 80 90 100 0.4
^ 0
h (X 100)

Figure 1. Distribution of diversity values generated
from 1000 bootstrap replicates from our sample of
razorback suckers collected in Lake Mohave, Arizona-
Nevada. Estimates for each replicate were calculated
from subsCl;mples of 10 individuals.

ability of resampling the same individual, thereby de-
creasing variance in distribution of expected h values. If
subsamples were drawn from a larger source popula-
tion, the diversity values even for the 1987 and 1989
year classes would be significantly different from those
of the source.
Although our estimate of diversity measures the rela-
tive frequency of haplotypes within a sample, it does not
provide an accurate indicator of change in the genetic
composition of hatchery stocks relative to natura] popu-
lations (see Allendorf & Ryman 1987 for a similar conclu-
sion relative to aUozyme diversity). In our simulations
we examined the behavior of h in high-diversity organ-
isms by resampling from a brocidstock in which every in-
dividual possessed a unique haplotype. When progeny
were derived from broodstocks of different size, mean h
values for progeny populations increased rapidly with
increased founder size, producing a curve that leveled
near h "'-'0.95 and broodstock sizes near 20 individuals
(Fig. 2). Therefore, in an organism with high diversity-
each individual possessing a unique haplotype- h is rel-
atively insensitive to broodstock size because the differ-
ence in diversity values from stocks generated by random
use of ova from 15-20 females will not be significantly
lower than those produced from 100 females.
The number of haplotypes (A) was also examined in
these simulations, and the increase in A was directly cor-
related with increase in broodstock size (Fig. 2). Given
that the progeny of 100 females will clearly possess
mor<; different genotypes than those produced from 15-
20, h misrepresents levels of variation maintained in the
broodstock, whereas A provides a much better indicator
of levels of genetic diversity.

Conservation Biology
Volume 10, No. I, February 1996

Dowling et al

80

60 Z
~
o
....,
e.
('D
('D
'"

40

^
.h
+A

'>
20 '-'

o

20

40 60

80 100

Broodstock size

Figure 2. Results of simulation depicting effects of
varying broodstock size on haplotype diversity (il)
and number of alleles (A). Broodstock sizes of 2, 5, 10,
15, 20, 25, 50, and 100 individuals were used.

The simulation results are consistent with our empiri-
cal data. Year classes in 1987 and 1989 exhibited similar
diversity values despite substantial differences in num-
ber of females used to generate each (Table 2). Thus,
variation in diversity was not due to the number of fe-
males used (17 in 1990 versus 55 and 14 for the 1987
and 1989 year classes, respectively). Given the relative
insensitivity of h to broodstock size, reduced diversity
in the 1990 year class indicates the difficulty in consis-
tently producing stocks that mimic a source population.

Factors Reducing Diversity in Hatchery Stocks

Reduction in diversity could be caused by extrinsic fac-
tors (hatchery effects) or intrinsic factors (differences in
fecundity/viability among females and their progeny), or
both. We cannot assess the role of extrinsic factors. They
should be consistent, however, because in this example
personnel, methods, and equipment at DNFH were rela-
tively constant from year to year.
The observed reduction in diversity might be ex-
plained by over-representation of some females in the
1990 year class. Initital comparison of the . average
number of ova per female recorded in DNFH annual re-
ports indicates no difference between the 1990 year
class ("-'0.94 X 105 eggs/female, n = 17) and 1982-1989
year classes ("'-'0.83 X 105 eggs/female, n = 590). These
statistics have little meaning, however, because large
fish have more ova than small ones, and the unrecorded
sizes of individual broodfish likely varied from 40 to 60
cm standard length.
On the other hand, survival from fertilized ova to
swim-up was only 11.1 % in 1990 (n = 17), substantially