I came across a problem in Cliff Pickover's Twitter feed. It is depicted in Figure 1.
![]() |
Figure 1 |
I came across a problem in Cliff Pickover's Twitter feed. It is depicted in Figure 1.
![]() |
Figure 1 |
I came across automorphic numbers in a tweet by Cliff Pickover. He pointed out that:
☛ \(376\) is an automorphic number, meaning a number whose square "ends" in the same digits as the number itself. \(376\) has the property that its cube and fourth power also end in the same digits.
I thought I'd investigate how many of these numbers there up to one million. It turns out that there aren't many. They are:
Surprisingly when we consider the cubes of numbers, the count increases substantially but the same numbers as for the squares reappear:
A033819 | Trimorphic numbers: \(n^3\) ends with \(n\). |
0, 1, 4, 5, 6, 9, 24, 25, 49, 51, 75, 76, 99, 125, 249, 251, 375, 376, 499, 501, 624, 625, 749, 751, 875, 999, 1249, 3751, 4375, 4999, 5001, 5625, 6249, 8751, 9375, 9376, 9999, 18751, 31249, 40625, 49999, 50001, 59375, 68751, 81249, 90624, 90625, ...
With fourth powers however, the count again is more modest and the same numbers reappear:
This property of these numbers continues indefinitely and as Wikipedia states:
There are four 10-adic fixed points of \( f(x)=x^{2}\), the last 10 digits of which are one of these:
Thus we see why all the automorphic number appear as they do, forming OEIS A003226. Apparently such numbers can also be called curious numbers or circular numbers.
A003226 | Automorphic numbers: \(m^2\) ends with \(m\). |
0, 1, 5, 6, 25, 76, 376, 625, 9376, 90625, 109376, 890625, 2890625, 7109376, 12890625, 87109376, 212890625, 787109376, 1787109376, 8212890625, 18212890625, 81787109376, 918212890625, 9918212890625, 40081787109376, 59918212890625, 259918212890625, 740081787109376, ...
Of course, automorphic numbers can exist in any base. For a given base \(b\), the number of \(b\)-adic fixed points is determined by 2^(number of distinct prime factors). Because 10 is the product of two distinct prime factors, it has \(2^2=4\) fixed points. Likewise with 6 and 12 (even though \(12=2^2 \times 3\), it has only two distinct prime factors). Of course, for prime numbered bases such as 2, 3, 5 etc. and perfect powers such as 4, 8, 9, 16 etc., there are only 2 fixed points and these are the trivial 0 and 1. Here is a permalink that will generate automorphic numbers in any base (up to 36) and for any power.
Applied to base 30 (that is comprised of three prime factors) it can be seen that there are \(2^3=8\) distinct 30-adic fixed points. Here are the 30-morphic numbers up to one million:
an integer sequence that starts with a positive integer \(a_0 \), with each subsequent term in the sequence defined by the recurrence relation:
Juggler sequences were publicised by American mathematician and author Clifford A. Pickover. The name is derived from the rising and falling nature of the sequences, like balls in the hands of a juggler. If a juggler sequence reaches 1, then all subsequent terms are equal to 1. It is conjectured that all juggler sequences eventually reach 1. This conjecture has been verified for initial terms up to one million, but has not been proved. Juggler sequences therefore present a problem that is similar to the Collatz conjecture, about which Paul Erdős stated that "mathematics is not yet ready for such problems".
Figure 1 shows the SageMath code (permalink) to determine the trajectory for any given number, along with the numbers of steps required and the maximum value reached.
![]() |
| Figure 1 |
Most numbers reach the value 1 quickly but others are more stubborn. Records are set as we move through the natural numbers and these numbers form OEIS sequence A094679. The sequence begins:
1, 2, 3, 9, 19, 25, 37, 77, 163, 193, 1119, 1155, 4065, 4229, 4649, 7847, 13325, 34175, 59739, 78901, 636731, 1122603, 1301535, 2263913, 5947165, 72511173, 78641579, 125121851, 198424189, ...
OEIS A094698 shows what these records are:
0, 1, 6, 7, 9, 11, 17, 19, 43, 73, 75, 80, 88, 96, 107, 131, 166, 193, 201, 258, 263, 268, 271, 298, 335, 340, 443, 479, 484
Comparing the two sequences we can see that there are 73 steps required for 193 to reach 1. The maximum value reached is a rather large during the trajectory is:
6743569603489758391265376070807357156339920158784377929096419715849060516985205368792190354996630779167466266586213526771780967700267133711091446786931423291036091166608223302792047793105565012490585915410391500762927066039966992101729450252321626382793545523711387059090
With such large numbers being involved, it's better to use a logarithmic scale for viewing the trajectory of a given number. For example, the trajectory of 1003 has 15 steps with maximum value 39526058. Here is its trajectory and Figure 2 gives a graphical representation: 1003, 31765, 5661392, 2379, 116035, 39526058, 6286, 79, 702, 26, 5, 11, 36, 6, 2, 1.
![]() |
| Figure 2: juggler trajectory of 1003 |
GeeksforGeeks gives the C++, C, Java, Python, C# and PHP code to generate the juggler trajectory for any natural number input.