on clinical cases, family histories, and some 
quantitative studies that indicate a partial 
hereditary basis, Lush (1950) analyzed data 
from 27 New Zealand herds, classifying cows 
as resistant, no mastitis up to 8 years, and 
susceptible, or those having mastitis. Using 
this all or none classification, about 85 per- 
cent of the daughters of susceptible dams 
were susceptible, whereas only 55 percent of 
the daughters of resistant cows were sus- 
ceptible. Intraherd regression of daughter on 
dam yielded a heritability of 0.387 0.32. 
Legates and Grinnells (1952) reported aherit- 
ability of 0.27 F 0.10, using data from 11 herds 
and classifications of streptococci or staphy- 
lococci with a leucocyte count of one-half mil- 
lion or more cells per milliliter in any quarter 
during any sampling period. 
A study by Young et al. (1960) is the most 
comprehensive to date. Their definitions of 
mastitis were (1) clinical - the percentage 
of months in lactation during which the cow 
had mastitis; (2) bacterial infection - scores 
of monthly milk samples; and (3) leucocyte 
score - computed by summing the logarithms 
of the monthly leucocyte counts. The estimates 
of heritability and repeatability are given in 
table 1. 
All these studies may be criticized for 
employing inadequate definitions of mastitis 
susceptibility. Since the scores utilized do 
not indicate sampling from a normal distri- 
bution, estimated progress in selection is 
not a very suitable criterion for use ina 
breeding program. Even so, the estimates 
are in close agreement and show important 
differences between animals, 
If we accept these heritability estimates and 
assume that resistance to mastitis is at least 
as heritable as milk production, what emphasis 
should be given in a breeding program? Even 
though several, including Hutt (1958) and 
Fredeen (1963), have advocated direct selection 
against susceptibility, little seems to have 
been done, The causes might be as follows: 
(1) To date we do not understand enough 
about the disease to construct a good index 
for ranking animals. For example, in the 
study of Young et al, (1960), the genetic cor- 
relation between clinical mastitis and bacterial 
infection of 0.23 to 0.29 indicates that selec- 
tion on either of these would have little effect 
on the other. The genetic correlation between 
leucocyte count and each of the other indicators 
of mastitis of 0.80 or higher would indicate 
this may be an effective criterion, However, 
since leucocytes are one of the most important 
defense mechanisms of the body, we may be 
selecting the most susceptible instead of the 
most resistant cows as parents. 
The work of Schalm et al. (1964) illustrates 
the point, as they found that mastitis developed 
in all quarters of cattle secreting cell-free 
milk when challenged with approximately 8,000 
Aerobacter aerogenes, but partial protection 
was afforded when the fore milk contained 
200,000-350,000 leucocytes per milliliter. 
Recent studies at Beltsville, Md., have shown 
a wide range (<30,000 to 10>million) of 
leucocytes shed by normal noninfected quarters 
(Smith and Shultze, 1964), whereas the number 
shed by infected quarters ranged from < 30,000 
to > 50 million, Leucocyte counts also appear 
to be affected by the type of bacteria infecting 
the quarter. 
(2) If selection is to be effective, it must be 
measured rather early in the reproductive 
life of an individual. The measure employed by 
Lush required cows to reach 8 years before 
they were considered resistant. Legates and 
Table 1.--Heritability and repeatability estimates for clinical mastitis, bacterial 
infection, and leucocyte count 


Method of evaluation 
Clima cailusmaisiGuliCslsperarcteeonare aveletoinakeseye 0.06 
Bacterialy intectionierae v.).mcse cee .18 
ewicoeyiceMecountieees save cles eholeiete .38 

Heritability 
Daughter-dam 
regression 
I+ |+ 1+ 


Repeatability 





Intracow 
correlation 
Paternal sister 
correlation 




0.18 0.31 + 0.06 
14 (2456s N06 
20 = 
