30
M. E. SOLOMON
subsequent count of fully fed larvae in May, and is calculated on the
basis of an assumed constant egg production.” They attribute it chiefly
to mortality of first-stage larvae, which they know to be both great and
variable. It is understandable that the higher values of k, are usually
followed in the same generation by relatively low values for larval
density, while the lower values of k, allow greater larval abundance.
The addition of the oensus data to the diagram also seems to confirm
emphatically that pupal predation is a promptly density-dependent
process, causing a proportionate mortality (k,) that rises and falls with
population density.
Provided the data are good enough, the method described by Varley
and Gradwell can be used to distinguish inverse, lagging and prompt
density relationships. However, it calls for more detailed observations
than the single factor method described by Morris (Zoc. cit.). The latter
requires only one annual census and an estimate of the mortality due
to the suspected key factor. Varley and Gradwell need mortality data
for each factor or group of factors to be represented by a k-curve. In
return for these extra demands, their method reveals the importance and
operational nature of as many factors as are represented by adequate
data. Thus, although they introduce their method as a type of key
factor analysis, its scope is wider than this. It has a good deal in
common with the analysis of survival rates undertaken by Morris
and his colleagues in their work on the spruce budworm, in which
survival ratios from successive mortalities in the life-cycle are multiplied
together and related to the change in numbers from each generation
to the next. The essence of Morris’ key factor method is to concentrate
on a single factor, irrespective of the way in which its action may turn
out to be related to density, consideration of other factors being
postponed until the time comes for more extended analysis by other
methods. Some of the ways forward from this stage are indicated by
Morris (1963) and Morris (ed., 1963). The method of Varley and Gradwell,
if reduced to its simplest form, would be the same as that of Morris.
One census each year would be required, together with observations to
provide a value for the mortality (k,) due to a suspected key factor.
The total mortality ( K ) would be estimated from the census figures and
a value for the egg output. Then all the mortality except k, would be
called k, ( =K - kl), and a graph would show what proportion of the
variation in K was due to k,.
3. By Life-table Analysis
In studying an insect population in the field one normally begins
without knowing which is the key factor (causing the main variation
in numbers from one generation to the next), and which is the regula-
M. E. SOLOMON
subsequent count of fully fed larvae in May, and is calculated on the
basis of an assumed constant egg production.” They attribute it chiefly
to mortality of first-stage larvae, which they know to be both great and
variable. It is understandable that the higher values of k, are usually
followed in the same generation by relatively low values for larval
density, while the lower values of k, allow greater larval abundance.
The addition of the oensus data to the diagram also seems to confirm
emphatically that pupal predation is a promptly density-dependent
process, causing a proportionate mortality (k,) that rises and falls with
population density.
Provided the data are good enough, the method described by Varley
and Gradwell can be used to distinguish inverse, lagging and prompt
density relationships. However, it calls for more detailed observations
than the single factor method described by Morris (Zoc. cit.). The latter
requires only one annual census and an estimate of the mortality due
to the suspected key factor. Varley and Gradwell need mortality data
for each factor or group of factors to be represented by a k-curve. In
return for these extra demands, their method reveals the importance and
operational nature of as many factors as are represented by adequate
data. Thus, although they introduce their method as a type of key
factor analysis, its scope is wider than this. It has a good deal in
common with the analysis of survival rates undertaken by Morris
and his colleagues in their work on the spruce budworm, in which
survival ratios from successive mortalities in the life-cycle are multiplied
together and related to the change in numbers from each generation
to the next. The essence of Morris’ key factor method is to concentrate
on a single factor, irrespective of the way in which its action may turn
out to be related to density, consideration of other factors being
postponed until the time comes for more extended analysis by other
methods. Some of the ways forward from this stage are indicated by
Morris (1963) and Morris (ed., 1963). The method of Varley and Gradwell,
if reduced to its simplest form, would be the same as that of Morris.
One census each year would be required, together with observations to
provide a value for the mortality (k,) due to a suspected key factor.
The total mortality ( K ) would be estimated from the census figures and
a value for the egg output. Then all the mortality except k, would be
called k, ( =K - kl), and a graph would show what proportion of the
variation in K was due to k,.
3. By Life-table Analysis
In studying an insect population in the field one normally begins
without knowing which is the key factor (causing the main variation
in numbers from one generation to the next), and which is the regula-