104 
 
 
site and genetical constitution at the two stations. 
Unless marked (b), the record for any given item 
may be assumed to approach the (a) category. That 
is to say, items not specifically designated as (b) are 
supposed to span the period from the earliest bloom 
on the earliest site to the latest bloom on the latest 
site. This implies, of course, that the observer ex- 
amines an equal variety of sites for each date for 
each year. Needless to say this ideal was not always 
carried out, but it usually was, especially for common 
species. 
  The choice between type (a) and type (b) is 
usually dictated by commonness. In some of the 
species, the population was too small for a record 
of the (a) category. Thus there is only one lilac 
bush [113] at the Sauk station. In such instances 
the record is marked (b) and an appropriate speci- 
fication, such as "one bush" or "one clump" is added.

  By and large, the (a) type is possible only for 
common species, while rare or locally uncommon spe- 
cies automatically assume the (b) type. The prac- 
tical point is to reach a decision, and stick to it, 
in species of intermediate abundance. 
  In aquatic and bog plants, whole ponds or bogs 
sometimes display uniform earliness or lateness de- 
pending on the presence or absence of springs. For 
example: 
 
 
Skunk cabbage [28] 
 
 
    with springs      without 
..-    3/20/45    about 4/10/45 
 
 
Marsh mazigold [711 .......  about 4/10/46 
 
 
4/15/46 
 
 
  For this reason a single specified pond or bog 
(phenology of the (b) type) is preferable for aquat- 
ics. A single water (Lake Chapman) was used for 
our Sauk records. 
                 SOURCES OF ERROR 
  The foregoing discussion describes certain particu- 
lar sources of error. There remain two more general 
questions: 
  (1) How much difference will occur as between 
two observers, looking for the same list of items in 
the same area, through the same years? 
   (2) How much difference arises from differing in- 
tensity of observation? 
          DIFFERENCE BETWEEN OBSERVERS 
  During parts of 1937, 1938, 1941, and 1942 Irven 
0. Buss recorded phenology at Faville Grove, Jeffer- 
son County, 20 miles east of Madison. This is in 
the eastern end of the area defined as eligible for 
the Dane station. These records were not seen until 
he returned from military service in 1945. Upon com- 
parison of his record with ours, 39 pairs of dates 
are found to be in common, covering 33 items dur- 
ing eight months of the year. Of the 39 pairs of 
dates, six are for birds, one for a mammal, and 32 
for plants. The disparities classify as follows: 
Identical with ours  __.--- 13 pairs        33 percent 
1-2 days difference ......... 12 pairs      31 percent 
3-4 days difference .......... 7 pairs      18 percent 
5-11 days difference .......  7 pairs       18 percent 
 
 
100 percent 
 
 
Ecological Monographs 
       Vol. 17, No. 1 
 
 
  The sum of all disparities is 101 days for 39 pairs, 
an average of 2.6 days each. 
  These disparities represent the error of observa- 
tion plus any actual differences that may exist be- 
tween the two localities. There is no available means 
for segregating these two components. 
 
             INTENSITY OF OBSERVATION 
                 Animal Behaviors 
  The writers have become convinced that most ani- 
mal behaviors in late winter are first exhibited by a 
few precocious individuals, and become general in 
the population by slow degrees. In 1946 some special 
records were kept to test this point. Many trained 
men, just returned from military service, enabled us 
to record not only the first occurrence of a behavior, 
but its gradual development. The following covers 
the combined observations of a dozen men afield 
daily on the University Campus and Arboretum. 
 
                                               Become 
Itemh    Species         Earliest Dates        Generql 
3     Cardinal song ......... Jan. 5ý 6, 9, 12, 13, 30  Feb. 17 
4     Pheasant crowing .----- Jasn. 10. 18, 30, Feb. 25 March 1 
5     Marshhawk arrives....Feb. 16           Feb. 24 
8     Chipmunk emerges.....Feb. 8, 10, 18, Mar. 7, 8 March 14 
23    Brown bat flying ...... Feb. 9         March 13 
39    Ruffed grouse drumsFeb. 9              March 16 
x     Garter snake out ..... Jon. 25, Feb 17 Eary March 
 
  The early dates for Items 3, 4, and 8 are believed 
to represent a slowly developing frequency; those 
for 5, 23, and 39 are believed to represent isolated 
early occurrences. In either case, a lesser intensity 
of observation would have yielded later "first dates," 
the degree of lateness and the selection of items 
depending on chance. 
  It is believed that this intensity-error in animal 
events diminishes rapidly as the season advances, 
and that it is never large in plant items, except in 
a few species to be discussed later. An intensity- 
error also probably exists in autumn animal items 
such as 316, 317, 319, 322, 324, and 326.. 
                   Bird Migration 
  Even if an event develops nearly simultaneously in 
a population, the chance of detecting its earliest oc- 
currence depends on the number of observers afield. 
To explore the magnitude of this factor, the bird 
arrival dates of one observer are compared with 
those of 10 observers at the Dane Station. 
  For 16 years beginning in 1913, A. XV. Schorger 
recorded the arrivals later published, as averages, in 
his "Birds of Dane County" (1929, 1931). He used 
only his own observations, and his skill as an orni- 
thologist is nationally known. He was able to make 
few field trips during week days. His dates repre- 
sent the best possible single-handed avocational effort. 
  Schorger's dates will now be compared with those 
gathered during the past decade by the writers and 
their collaborators: 
 
 
ALDO LEOPOLD AND SARA ELIZABETH JONES 
 
 
39 pairs