Here is why divisibility determines when the entry does not increase by 1 along a column in Wouter Meeussen's (tab 2) :
Meeussen's (tab 1) of the integer partitions of n=7 gathered first by length and then by first part is a convenient example to illustrate the role of divisibility in (tab 2). The count of the groupings for n=7 is {1,3,3,3,2,1,1}, while (tab 2) indicates that the count of the groupings for n=6 is {1,3,3,2,1,1}. The question is, why is the number of groupings for length=2 and length=3 the same for n=7 as for n=6?
Consider length=2:
The partitions of 6 are {5,1}, {4,2}, {3,3}.
The partitions of 7 are {6,1}, {5,2}, {4,3}.
The partitions of 8 are {7,1}, {6,2}, {5,3}, {4,4}.
It is evident that when n is even, the partition {n/2, n/2} occurs, along with the partitions {n-1, 1}, {n-2, 2}, ..., {n/2 + 1, n/2 - 1}.
But when n is odd, there is no partition {n/2, n/2}. So the only partitions of length=2 are {n-1, 1}, {n-2, 2}, ..., {(n+1)/2, (n-1)/2}.
The value (n-1)/2 for odd n is the same as the value m/2 for even m = n-1. This is why the number of groupings of partitions of length=2 must be the same for odd n as for even n-1, and it is why the entry does not increase by 1 along the given column of (tab 2).
And for length=3:
The partitions of 6 are {4,1,1}, {3,2,1}, {2,2,2}.
The partitions of 7 are {5,1,1}, {4,2,1}, {3,3,1}, {3,2,2}.
The partitions of 8 are {6,1,1}, {5,2,1}, {4,3,1}, {4,2,2}, {3,3,2}.
The partitions of 9 are {7,1,1}, {6,2,1}, {5,3,1}, {5,2,2}, {4,3,2}, {3,3,3}.
The partitions of 10 are {8,1,1}, {7,2,1}, {6,3,1}, {6,2,2}, {5,3,2}, {4,4,2}, {4,3,3}.
It is evident that when n == 0 mod 3, its partitions of length n=3 have first parts n-2, n-3, ..., n/3.
When n == 2 mod 3, its partitions of length n=3 have first parts n-2, n-3, ..., (n+1)/3.
When n == 1 mod 3, its partitions of length n=3 have first parts n-2, n-3, ..., (n+2)/3.
This means that any n == 1 mod 3 has the same number of groupings of length=3 as n-1, but any n == 0 mod 3 or n == 2 mod 3 each has 1 more grouping of length=3 than n-1.
This is why the entry does not increase by 1 only for n == 1 mod 3 along the given column of (tab 2).
The same can be demonstrated for any length=k :
n == 0 mod k has partitions with first parts n-k+1, n-k, ..., n/k.
n == 1 mod k has partitions with first parts n-k+1, n-k, ..., (n+k-1)/k.
n == 2 mod k has partitions with first parts n-k+1, n-k, ..., (n+k-2)/k.
etc.
It follows that each n has 1 more grouping of length=k than n-1, EXCEPT for the case n == 1 mod k, which has the same number of groupings of length=k as n-1.