Suppose is a positive
integer in the list . This is
equivalent to , which is equivalent to since , , and are non-negative.
Fix an integer satisfying
and assume that is an integer satisfying both and . That means is the first digit past the decimal
point in the decimal expansion of . That is between and means the integer part of is . Therefore, is at least more than and at most more than . Since is between two consecutive
perfect squares, it is not a perfect square, which means is irrational. Putting this
together, we get that where the strict inequalities
are justified by the irrationality of . We are also assuming that is a positive multiple of , which means there is some positive
integer such that .
The quantities in the chain of inequalities above are all positive,
which means Expanding
and substituting , we get
We will now count the number of integers that satisfy the
inequalities above.
Since the integer satisfies
, we must have that
.
Therefore, is between the
consecutive integers and
, possibly equal to the
smaller of the two but not equal to the larger of the two. Since is an integer that is greater than
, we
conclude that
Similarly, implies
that . This means is
between the consecutive integers and , possibly equal to the
larger of the two but not equal to the smaller of the two. Since is an integer that is smaller than
,
we can conclude that . Combining this inequality with the inequality
from earlier, we now have that The number of integers that satisfy this inequality is
We have now shown the following: For each integer satisfying , there are at most integers in the list with the
property that . There are
integers in the list above and
only ten different values that the function can output. Since , there is no possibility other
than that takes every possible
value exactly
times.
To get an idea of how to proceed, we first discuss the results of
parts (a), (b), and (c).
In part (a), we saw that for values of between and , the function takes on every possible value the same
number of times (zero) with the exception of and , which occur one extra time each. For
values of between and , the function takes on every possible value the same
number of times (one) with the exception of and , which occur one extra time each.
Similarly, in part (b) we observed that takes on every value the same number of
times, with the exceptions of ,
, , and when ranges between and as well as between and .
The result of part (c) was that if is a multiple of , then takes on every possible value the
same number of times (with no exceptions) as ranges over the integers strictly
between and .
The patterns in part (a) and (b) continue, and there are similar
patterns for other ranges between perfect squares. A bit more precisely,
if is a positive integer, then as
ranges over the integers between
and , takes on every possible value the
same number of times, possibly with some exceptions that occur one extra
time each. These exceptional values depend only on the remainder when
is divided by .
Even more precisely, the following five claims are true. The claims
are numbered to correspond to remainders after division by .
Claim 0: Suppose for
some integer . As ranges over the integers strictly
between and , takes every value from through exactly times.
Claim 1: Suppose for
some integer . As ranges over the integers strictly
between and , takes the values and a total of times each and takes every other
value times.
Claim 2: Suppose for
some integer . As ranges over the integers strictly
between and , takes the values , , , and a total of times each and takes every other
value times.
Claim 3: Suppose for
some integer . As ranges over the integers strictly
between and , takes the values , , , , , and a total of times each and takes every other
value times.
Claim 4: Suppose for
some integer . As ranges over the integers strictly
between and , takes the values , , , , , , , and a total of times each and takes the values and a total of times each.
In part (c), we showed that Claim 0 is true. Before proving Claims 1,
2, 3, and 4, we will answer the actual question which was to determine
how many times each digit occurs in the list Because
is a perfect square, , so we can instead count
the number of times each digit occurs in the slightly different list
There are perfect squares
among the integers from through
, so there are occurrences of in the list that come from perfect
squares.
For every other integer in the
list , there
are unique integers and with and
with the property that .
Suppose is a non-zero digit.
For fixed and , as we let take the values strictly between and , Claim tells us whether the digit occurs either times or times in that range. We are
considering values of satisfying
for
every pair with and . For every pair of the form
, we get occurrences of (see Claim 0). For every pair of the
form , we always get occurrences or we always get occurrences of , depending on what Claim 1 says about
the digit . Continuing with this
reasoning, the number of occurrences of in the list can be
expressed in the form where is the
number of values of for which
Claims 0 through 4 say there are
occurrences of , and is the number of values of for which Claims 0 through 4 say there
are occurrences of .
For example, for , there are
occurrences when , , and , and there are occurrences when and . Therefore, with , we have and , so the number of times that occurs in the list is .
Using Claims 0 through 4, the table below summarizes the count for
the remaining non-zero digits. In the first column, the digit appears, in the second column the value
of appears, in the third column,
the value of occurs, and in the
fourth column, appears,
which is the number of times
appears in the list.
Finally, for , we already
have occurrences of in the list coming from the perfect
squares. Otherwise, we can use the exact same technique as above to
count the number of s in the list
coming from integers that are not perfect squares. For , and , so there are occurrences of the
digit in the list .
We will now prove Claims 1 through 4.
Suppose is a non-negative
integer, is , , , , or , and that is a digit between and inclusive. We wish to count the number
of integers such that and .
Specifically, we want to show that this number is either or depending on and in the way delineated in Claims 0
through 4.
The condition is equivalent to , and this combined with (note that is not a perfect square and that and are consecutive integers) is
equivalent to
Since everything is non-negative, everything can be squared to get the
equivalent chain of inequalities
After some expansion, we have We
are interested in how many integers satisfy the chain of inequalities
above. Since is an
integer, the number of integers
satisfying the inequalities above is the same as the number of integers
that satisfy and after
combining fractions on the left and right, we get
Suppose
and
are both strictly between and
. Then the integers satisfying are precisely the integers through . More generally, we will show that
whether there are or integers depends on whether or not
the quantities and are between
the same pair of consecutive integers.
To give an idea of how the argument will proceed, consider the
situation when and . Then ,
but .
Thus, becomes The integers through satisfy these inequalities,
for a total of integers.
Observe that This means the difference
between and
is
less than , which leads to the
following two possibilities.
Possibility 1: There is an integer so that In this case, the integers satisfying are , , , and so on through for a total of integers.
Possibility 2: There is some integer with the property that
In this case, the integers satisfying are through , for a total of integers.
We have now shown that there will be integers satisfying exactly when there is an integer
strictly between and .
Multiplying through by , this is
equivalent to there being a multiple of strictly between and
The table below contains the values of for every and . The columns after the
first are indexed by a value of
from through from left to right and the rows after
the first are indexed by values of from through from top to bottom. The cell in the row
corresponding to and the column
corresponding to contains . Cells are highlighted if there
is a multiple of between the
value in the cell and the value in the cell immediately to its
right.
/ |
0 |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
0 |
0 |
1 |
4 |
9 |
16 |
25 |
36 |
49 |
64 |
81 |
100 |
1 |
0 |
21 |
44 |
69 |
96 |
125 |
156 |
189 |
224 |
261 |
300 |
2 |
0 |
41 |
84 |
129 |
176 |
225 |
276 |
329 |
384 |
441 |
500 |
3 |
0 |
61 |
124 |
189 |
256 |
325 |
396 |
469 |
544 |
621 |
700 |
4 |
0 |
81 |
164 |
249 |
336 |
425 |
516 |
609 |
704 |
801 |
900 |
The highlighted cells correspond to pairs for which is less than a multiple of and is greater than that
multiple of . As noted above,
these correspond to the pairs
for which integers satisfy
. For , we see no values of , which gives another proof of Claim 0.
For , the values of for which there are integers are and , which proves Claim 1. Continuing,
the data in the table above verifies Claims 0 through 4.