This work is licensed by its author, Gary Croft, under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
"This sieve is the source of two efficient prime number factorization algorithms and completely deconstructs the twin prime sequence." – Philip Gervasse Jackson, citing the Prime Spiral Sieve in his Simplicity Instinct: Why Prime Numbers are Elusive!
When the set of natural numbers not divisible by 2, 3 or 5 is spiraled within a modulo 30 factorization wheel (see graphic, below), where 1 = 12°, all twin primes and twin prime candidate pairs of the form n, n + 2 (with the exception of [3,5] and [5,7]), are arrayed along three distinct "distribution channels," i.e., potential twin prime pairs are separated by 24° and are parsed equally along three sets of paired radii: [12° ↔ 348°] ... [132° ↔ 156°] ... [204° ↔ 228°] ... in intervals of 30:
Translated into modular arithmetic, modulo 30 for the three twin prime candidate channels (A, B and C, for convenience) distribute as follows, and in the sequences shown:
A: [n & n+2] ≡ [1, 29] (mod 30)
B: [n & n+2] ≡ [11, 13] (mod 30)
C: [n & n+2] ≡ [17, 19] (mod 30)
And in digital root terms they translate to:
A: 1 + 29 = 30 = dr(3)
B: 11 + 13 = 24 = dr(6)
C: 17 + 19 = 36 = dr (9)
(Toward the bottom of this page, we learn why the {3,6,9} triangulation is significant when we discuss the "Trinity of Triangles" and reveal the 'magic square' reflecting {3,6,9} hidden in plain sight betwixt and between the twin prime distribution channels.)
[Note: In our 'Foundations' section, we stated that the set of natural numbers not divisible by 2, 3 or 5 can be expressed in the (modulo 30) form: 30n+1, 30n+7, 30n+11, 30n+13, 30n+17, 30n+19, 30n+23, 30n+29, once factor-sieved leaving all primes >5. Sub-sorting (to use a spreadsheet term) the prime pair candidates in the twin prime distribution channels can be expressed as (30n+11, 30n+13), (30n+17, 30n+19), (30n+29, 30n+31) ... which narrows the factor-sieving field from 8/30 to 6/30, or 20% of all natural numbers. We've documented this sequence on the Online Encyclopedia of Integer Sequences (OEIS Sequence Number A230462) where, incorporating a suggestion from Omar E. Pol, we've defined it as Numbers congruent to {1, 11, 13, 17, 19, 29} mod 30.]
When we consolidate the 6 radii used to construct the twin prime distribution channels into a single number line, we can completely isolate twin prime candidates, pairing them in tandem in a deterministic sequence, viz.: 11 {+2+4+2+10+2+10} {repeat ... ∞}. For example, six of the first seven such pairings are twin primes: [11, 13] [17, 19] [29, 31] [41, 43] [47, 49 (7 x 7)] [59, 61] [71, 73]. Below is a matrix showing the distribution of twin prime sets (both candidates and confirmed prime pairs) from [11, 13] to [569, 571]:
The first three twin prime sets in this matrix configuration are an anagram, in angular terms (in a 30-sectioned circle like the Prime Spiral Sieve, where 1 = 12°), to the first three primes, thus: [11, 13] [17, 19] and [29, 31] respectively translate to:
11 + 13 = 24 ... 24° = 2
17 + 19 = 36 ... 36° = 3
29 + 31 = 60 ... 60° = 5
[You'll note that the sum of the intervals for each row of twin prime candidates = 30 = 2 x 3 x 5; and interestingly the product of the sums of the first three twin prime candidate sets = 24 x 36 x 60 = 51840, while 51840/360 = 12^{2}]
All composite numbers in the twin prime distribution channels can be accounted for by the eight-fold algorithmic progressions shown below, where we first show factorization by multiplication, then follow with factorization by division. [Note: The algorithms shown here are a sub-set of the chordal progressions discussed at some length earlier that factored all composite numbers in the array populated by the set of natural numbers not divisible by 2, 3 and 5, as opposed to only factoring those impacting twin prime candidates, demonstrated below.]:
Another way to demonstrate twin prime candidate factorization is by sequential division. As you can see below, the left-most column consists of 1 and members of the twin prime distribution channels in sequence, thus delimiting the factorization to only twin prime candidates. Essentially, we've taken the sequenced elements in the twin prime distribution channels (30n+11, 30n+13), (30n+17, 30n+19), (30n+29, 30n+31) and divided them by the set of natural numbers not divisible by 2, 3 or 5: 30n+1, 30n+7, 30n+11, 30n+13, 30n+17, 30n+19, 30n+23, 30n+29. A more concise way to express this is that we divide numbers congruent to {1, 11, 13, 17, 19, 29} mod 30 by numbers congruent to {1, 7, 11, 13, 17, 19, 23, 29} mod 30. All modulo n ≡ 0 results generated (except when a number is divided by itself) indicate factors (arrayed horizontally, hi-lited in pink) of n (arrayed vertically, hi-lited in yellow, except for twin primes hi-lited in bright blue):
From what we learned above about segregating twin prime candidates, we can demonstrate that they compile additively in perfect progression, completing an infinite sequence of circles (multiples of 30 and 360):
.We've multiplied the twin prime candidate pair sums by 15° to show the equivalent arc of rotation of each pair in a modulo 90/24-element per rotation spiral, where 1 = 15° and 24 x 15 = 360°. Thus, when you divide the products by 360°, you get circular equivalents, e.g., 11+13=24; 24x15=360; 360/360=1 (thus, the 1st pair in the first thirt is equivalent to one 360° rotation).
A thirt, in case you're wondering, is a useful unit of measure when discussing intervals in natural numbers not divisible by 2, 3 or 5. A thirt, equivalent to one rotation around the Prime Spiral Sieve is like a mile marker on the prime number highway. If we take the Modulo 30 Prime Spiral Sieve and expand it to Modulo 360, we see that there are 12 thirts in one complete circle, or 'cirque' as we've dubbed it. Each thirt consists of 8 elements. The first two thirts are {1,7,11,13,17,19,23,29} and {31,37,41,43,47,49,53,59}. The last element in thirt1, 29, and the 1st element in thirt2, 31, are rooted in -1 and 1, respectively: where -1+30=29 and 1+30=31. These are twin primes, followed by an infinite sequence in intervals of 30 of twin prime candidates, that serve as the center stripes of the prime number highway, and thus: {-1,+1} +30 {29,31} +30 {59,61} +30 {89,91} +30 {119,121} +30 {149,151} +30 {179,181} +30 {209,211} +30 {239,241} +30 {269,271} +30 {299,301} +30 {329,331} +30 {359,361} completes cirque1, albeit, paradoxically, 361 is the start of cirque2. You can say that {359,361} straddle the 1st and 2nd cirques exactly at 360°, just as {719,721} straddle cirque2 and cirque3 at 720 (or 360x2) ... This is congruent with the fact that all repeating patterns at the digital root level, including the digital roots of Fibonnaci numbers and Lucas numbers indexed by our domain as well as the digital roots of the set constituting our domain, achieve in toto end-to-end symmetrical resolution at 360° intervals, which, of course, is aesthetically satisfying.
Here's a graphic that illustrates these relationships:
All prime numbers among the index numbers and the Prime Root Set have been hi-lited in blue. Of the 24 index numbers that are prime, 22 point to elements of the Prime Root Set that are also prime; and the two that aren't, 161 (7x23) & 221 (13x17), are both semi-primes with factors that sum to 30 (and, interestingly, the 30th index # points to 109, a prime number that–like 89–has a decimal expansion based on the Fibonnaci sequence). You can see from the graphic above that the digital roots of the Fibonacci numbers indexed to our domain (Prime Root Set) repeat palindromically every 32 digits (or 4 thirts) consisting of 16 pairs of bilateral 9's. The digital root sequence of the Prime Root Set, on the other hand, repeats every 24 digits (or 3 thirts) and possesses 12 pairs of bilateral 9's. The entire Prime Root sequence end-to-end covering 360° has 48 pairs of bilateral 9's. And finally, the Prime Root elements themselves within the Cirque, consisting of 96 elements, has 48 pairs of bilateral 360's. Essentially, the prime number highway consists of infinitely telescoping circles ... Also note, the digital roots of the Prime Root Set as well as the digital roots of Fibonnaci numbers and Lucas numbers (the latter not shown above) indexed to it all sum to 432 (48x9) in 360° cycles. The sequence involving Fibonacci digital roots repeats every 120°, and has been documented by the author on the Online Encyclopedia of Integer Sequences (Sequence #A227896): Digital root of Fibonacci numbers indexed by natural numbers not divisible by 2, 3 or 5 (A007775).
[Note: In some circles (no pun intended) the number 432 (as in 432Hz) has been declared the "frequency of the cosmos" to which the "Music of the Spheres" is tuned. A growing number of musicians and musicologists support tuning musical instruments to A=432Hz (or Pythagorean A) rather than the standard concert pitch of A=440Hz. They claim that 432Hz, with its 12-scale octave overtones compared to 440Hz's 8 octave overtones, is harmonically richer and cleaner in sound (The author, who plays American World Music on a flamenco guitar, will attest to this.). It's clear, upon analyzing the cirque de primes, that the digital roots sum to 432 in 360° cycles encompassing 12 thirts. Common sense, if not mathematical rigor, would suggest that there is some correspondence to 432hz's 12-scale octave overtones (and in fact, the Prime Spiral Sieve is featured prominently on this Latin American website's page titled "Fundamental Theorem of Arithmetic Applied Musically to the Zodiacal Circle"). Go to YouTube.com, search on "432," and you'll be impressed by the number of videos featuring this number. One more tantalizing factoid: 432^{2} closely approximates the speed of light in miles per second: 186,282.397 mps vs. 186,000.]
We get another cirque perspective when we sum the intervals between the twin prime candidate pairs, as matrixed below:
Another way to demonstrate the progression of a cirque is with a simple stair-step function, where x = 0...+15 ... repeat ... n, thus: x^{2} - (x - 1)^{2} = n. The resulting 12 steps are all n ≡ 29 (modulo 30) and their intervals, encompassing cirque1, sum to 360. And if you add 2 to each step you create the twin prime distribution channel rooted in n ≡ 29 (modulo 30) and n ≡ 1 (modulo 30), the 1st prime pair in this sequence being [29,31]:
0^{2} - 1^{2} = -1
15^{2} - 14^{2} = 29
30^{2} - 29^{2} = 59
45^{2} - 44^{2} = 89
60^{2} - 59^{2} = 119
75^{2} - 74^{2} = 149
90^{2} - 89^{2} = 179
105^{2} - 104^{2} = 209
120^{2} - 119^{2} = 239
135^{2} - 134^{2} = 269
150^{2} - 149^{2} = 299
165^{2} - 164^{2} = 329
180^{2} - 179^{2} = 359
Now lets zoom in to examine the first 90° partition (encompassing 3 thirts) of our Cirque de Primes, that starts with 1 and ends with 89. By juxtapositioning the 'Magic Mirror Matrix' against this range (see graphic, below), beautiful hidden symmetries are revealed at the digital root level when matrix factorized. This symmetry repeats every 90 degrees, repeatedly incrementing all 24 elements +90 (jumping from 1 thru 89 to 91 thru 179 then +90...n) while the digital roots never change. And thus, this 'geoarithmetric' object, telescoping to infinity, represents the factorization sequencing that accounts for all composite numbers extant in the set of natural numbers not divisible by 2, 3 or 5, leaving all prime numbers greater than 5. For an in depth discussion of the 'Magic Mirror Matrix,' including stunning facts about the number 89 (which, among other things, has the Fibonnaci sequence embedded in the decimal expansion of its reciprocal), click here: Magic Mirror Matrix.
Parsing Twin Primes by their digital roots (in context with natural numbers not divisible by 2, 3 or 5) reveals extremely interesting patterns. Study the matrix below, and you'll discover that it's a beautiful construct of perfectly balanced, multi-directional, 'logic tight' symmetries:
Note 1: We've documented the three twin prime digital root dyad sequences shown above and discussed below on the Online Encyclopedia of Integer Sequences (Sequences numbered A232880, A232881 and A232882). Here are the links
A232880: Twin primes with digital root 2 or 4
A232881: Twin primes with digital root 5 or 7
A232882: Twin primes with digital root 8 or 1
Note 2: As indicated, the pink squares marked with x's denote elements with terminating digits = td(5) (their values neither shown nor added into the row sums, above). Although they are not members of our defined set, they are nonetheless interesting. For one thing, the first six td(5) elements sum to 360, as follows: 25 + 35 + 55 + 65 + 85 + 95 = 360. For another, they are always adjacent to modulo 30 "Isolated Primes," discussed below.
Note 3: The first six Isolated Primes, starting with 23, also compile to 360, as follows: 23 + 37 + 53 + 67 + 83 + 97 = 360.
Note 4: Seven is the only number arrayed along the two Isolated Prime radii that isn't 'isolated,' given it is twinned with number 5. Were we to include 7 in our summing of isolated primes, above, the result would be 367, which is itself prime and the sum of seven prime numbers.
Note 5: The first 12 elements that are not td(5), consisting of 11 consecutive prime numbers plus 7^{2}, also sum to 360, as follows: 11 + 13 + 17 + 19 + 23 + 29 + 31 + 37 + 41 + 43 + 47 + 49 = 360.
Note 6: Every row possesses "sum symmetry" involving three identical sums. The first row, for example, gives us three sums (working outside-in) totaling 36: 11 + 25; 13 + 23; and 17 + 19 = 36 (and given 36 is 360/10, it comes as no surprise that the 10th row has three sums totaling 360). The 2nd row has three sums totaling 72; the 3rd row has three sums totaling 108, etc. Each successive row thus increments +36, and the row sums accumulate in overlapping cycles of 360 and 432 (the curious relationship between these two numbers discussed elsewhere on this site).
When we described the Digital Root Parsing Matrix above as 'logic tight,' we were not overstating it, as evidenced by the fact that, when elements with a terminating digit of 5 are included, every element in the first 49 rows of the matrix can be coupled with (added to) another element in the matrix to equal 900 (with the paired elements spatially located in "equal and opposite" positions). [Note: Please keep in mind that it is completely arbitrary of us to fucus on the '900 Block' given there are an infinite number of blocks in an infinite number of incremented configurations within the bounds of the infinitely expandable matrix (e.g., from 47 in Row 3 to 313 in Row 17 forms a '360 Block' and 11 in Row 1 to 1789 in Row 99 (not shown above) would create an '1800 Block').] Obviously, each pair totaling 900 must be comprised of elements producing digital root sums that decompose to dr(9). We've dubbed the 48 dyads consisting of two prime numbers totaling 900 in this block '900 Block Primes' (listed below):
The Digital Root Parsing Matrix is structured such that any rectangular configuration with 'n' number of columns and 'n' number of rows will produce lateral dyadic sum symmetry, determined by the three dr(9) sum variations: {2,7}, {4,5} and {8,1} (Note: To calculate the bilateral sums for a row, multiply the row's index number times 36.). Here are two examples involving complete (6-element) rows: Take the 17 rows from 2 thru 18 and you'll produce 51 combinations summing to 360; or take the 5 rows from 19 thru 23, which produces 15 combinations summing to 756 (To calculate the number of combinations for any and all rectangles configured from complete rows, multiply the given number of rows times 3.). Now lets carve out a rectangle that doesn't utilize complete rows: we create a 4-column/6-row rectangle with the four corners = 11, 19, 101 and 109 (all prime), and find that all cross-lateral sums equal 120.
What does this object tell us about twin primes? First, we see that twin primes fall into three categories when parsed by their digital roots: {2,4}, {8,1} and {5,7} (2x4, 8x1 and 5x7 all equal dr(8), while 2+4=dr(6); 8+1=dr(9) and 5+7=dr(3): the three legs of the {3,6,9} triangulation which is a member of the Trinity of Triangles, discussed toward the bottom of this page. Next we learn that every twin prime candidate set wihin the '900 Block' has a twin prime candidate pair 'in opposition' creating two sums of 900. Of those falling into this category within the block above, our favorite consists of four sets of twin primes, thus:
Employing the logic of digit sum arithmetic, we deduce that there are a finite number of factorization combinations that correlate to each of the six digital roots, for a total of 6x6 = 36 dyads ... and isn't it beautiful that 10x36 = 360...
The diagram below shows the first 26 twin prime candidates (13 pairs) from the Digital Root Parsing Matrix including the 6 composite numbers falling within this range (from {11,13} to {131,133}). For each composite number we show both its digital root and natural factors.
As we've discussed previously, every pair of twin prime candidates has one element each from the {1,4,7} and {2,5,8} alternating triangulations (and as indicated, all squares distribute to {1,4,7}–never to {2,5,8}).
Most notably, perhaps, digital root parsing reveals the clustering of twin prime candidates into "six packs," i.e., groupings into three prospective prime pairs demarcated by the isolated prime radii. [Note: The only twin prime set in our domain not in a six-pack {17, 19} sums to 36, which, as 1/10th of 360, seems poetic. If we were to decrement {17, 19} in two -18 steps, the result would be 17 + 19 - 1 + 1 - 17 - 19 = 0, which makes logical sense (from zero to 180 being consistent with the set incrementation, shown below).] The first eleven six packs follow (and you'll find the first three of these in the graphic above). And as we alluded, the sets sequence in +180 increments, thus:
11+13 + 29+31 + 47+49 = 180
41+43 + 59+61 + 77+79 = 360
71+73 + 89+91 + 107+109 = 540
101+103 + 119+121 + 137+139 = 720
131+133 + 149+151 + 167+169 = 900
161+163 + 179+181 + 197+199 = 1080
191+193 + 209+211 + 227+229 = 1260
221+223 + 239+241 + 257+259 = 1440
251+253 + 269+271 + 287+289 = 1620
281+283 + 299+301 + 317+319 = 1800
311+313 + 329+331 + 347+349 = 1980
If you think of the above as an 11-row/6-column matrix (excluding the row totals), you'll discover that its 66 elements form 33 dyads that sum to 360 (e.g, 11+349 = 360). Each column stair-steps in increments of 30 and each row, working from the outside in, has three identical sums (e.g., the first row has three dyads that sum to 60 each). Also, we find that the repeating digital root sequences reflect the Trinity of Triangles {3,6,9} {1,4,7} {2,5,8}:
2+4+2+4+2+4: dr(2) + dr(4) = dr(6)
5+7+5+7+5+7: dr(5) + dr(7) = dr(3)
8+1+8+1+8+1: dr(8) + dr(1) = dr(9).
Switching metaphors, below we represent our repeating 'six-pack' clusters as dominoes. The first row of dominoes reveals patterns shared by all three of the 3-pair twin prime candidate configurations, namely the terminating digit and modulo 30 values array identically across the three dyads parsed by digital roots. The second row of dominoes reveals the repeating patterns by digital root and then by digital root with the terminating digit values excluded. [Note: Every category below possesses 'equal and opposite' sum symmetry. For example, the modulo 30 configuration has 9 lateral sums of 30 each.]
It is evident studying the above that the difference sequence between elements in like positions in the 3-pair clusters is 90. We know from digit sum arithmetic that when you increment +90, both the original elements and the +90 incremented sums will possess the same terminating digits, digital roots, n(modulo 30) results, as well as the same digital roots when terminating digit values are excluded. For example, take the twin prime candidate pairs {89, 91} +90 = {179, 181}:
89: dr(8); td(9); 29(mod 30); dr(n-td) = 8
89+90 = 179
179: dr(8); td(9); 29(mod 30); dr(n-td) = 8
91: dr(1); td(1); 1(mod 30); dr(n-td) = 9
91+90 = 181
181: dr(1); td(1); 1(mod 30); dr(n-td) = 9
The dominoes also tell us that:
The lower-valued (left-hand) member of a prime pair will always possess one of three terminating digit ↔ modulo 30 combinations:
td(1) ↔ 11(mod 30)
td(9) ↔ 29(mod 30)
td(7) ↔ 17(mod 30)
The higher-valued (right-hand) member of a prime pair will always possess one of three terminating digit ↔ modulo 30 combinations:
td(3) ↔ 13(mod 30)
td(1) ↔ 1(mod 30)
td(9) ↔ 19(mod 30)
And lastly, twin prime sets that are: n ≡ 29(mod 30) + 2 = n ≡ 1(mod 30) decompose to: dr(n-td) = 2, 3, 5, 6, 8 or 9, while all other twin primes sets [namely n ≡ 11(mod 30) + 2 = n ≡ 13(mod 30) and n ≡ 17(mod 30) + 2 = n ≡ 19(mod 30)] decompose to: dr(n-td) = 1, 4 or 7.
When we convert the Twin Prime Distribution Channels and/or the paired radii whose angles sum to 360° into digital root multiplication matrices, the results are interesting–to say the least. At the digital root level there are six interrelated matrices of special interest to us: three represent the twin prime distribution channels; the fourth captures the two 'isolated' radii of the Prime Spiral Sieve, i.e. where n ≡ 7 (modulo 30), at 84°, and n ≡ 23 (modulo 30), at 276° (isolated, because, with the exception of 7, which is twinned with 5, the closest proximity they can be to other primes is plus or minus 4, and therefore they can never be twin primes); the fifth and sixth matrices complete the set of four derived from radii whose angles = 360° and in modulo 30 terms sum to 30.
Here's a diagram showing the derivation of the six matrices:
Each of the six matrices draws 24 elements from the first 96 elements of our domain (starting with 1 and ending with 359; you'll find a deep analysis of these 96 elements under the heading 'Cirque de Primes,' above ...). When matrix multiplied, each sequence produces sixteen identical 36-element squares comprised of six 6-element sequences. The matrix derived from the twin prime distribution channel where n ≡ 1 (modulo 30), at 12°, and n ≡ 29 (modulo 30), at 348°, is pictured directly below. It has sequences identical to those generated when you take a digital root multiplication matrix (aka Vedic Square) and remove all 3's, 6's and 9's (This "Imaginary Square," as we dubbed it, is discussed in more detail under prime factorization).
The matrices for twin prime distribution channels B & C follow:
Our next matrix is derived from the Isolated Prime Distribution Radii:
And lastly we present the third and fourth matrices derived from paired diagonals whose angles sum to 360°:
Each of the digital root multiplication matrices produced by the six channels consists of what are known in mathematics as 'Orthogonal Latin Squares' (defined in Wikipedia as "an n x n array filled with n different symbols, each occurring exactly once in each row and exactly once in each column" ... in our case every row and column of the repeating 6x6 matrices possesses the six elements: 1, 2, 4, 5, 7, 8 in some order). Also, the sum of the multiplicative digital roots = 108 x 24 = 2592 = 432 x 6.
[Note: Channels A, D, E and F combined represent the set of natural numbers not divisible by 2, 3 and 5, the first 24 elements of which form the basis of the Magic Mirror Matrix.]
The graphic below illustrates the transformative relationships between the matrices employing their primary building blocks (one of the sixteen identical 6 x 6 (36 element) Latin Squares that constitute each matrix):
Focusing on just the twin prime distribution channels, we see the relationships shown below [and, directly above, we show that two of the channels (B & C) transform bi-directionally by rotating 180° around one of their principal (lower-left to upper-right) diagonal axes]:
In our discussion under Cirque de Primes, above, we noted that the digital root of Fibonacci numbers indexed by natural numbers not divisible by 2, 3 or 5 form a 32-beat palindromic sequence that repeats every 120° and sums to 432 every 360°. When matrix multiplied this sequence forms a pyramidal mandala when stared at 'for a spell' ... a spell cast by reflectional symmetries within symmetries within symmetries. If there were room here, we'd expand this matrix times 3 to encompass 360°. Its 96 palindromic rows and 96 palindromic columns would all equal the magic number 432 ... (Also consider that the digital root of 96 = 6 and thus each row and column when matrix multiplied = 6 x 6 = 36 = 360/10, while the digital root of 36 = 9, which is equivalent to 0 in digit sum arithmetic, i.e., Alpha and Omega; Being and Nothingness.) Go figure ...
When we index the Fibonacci digital roots matrixed above by the twin prime digital root parsing method described earlier, where we parse by digital root dyads {2,4}{1,8} and {5,7}, bilateral sum symmetries are revealed:
To complete the picture (the 4th dimension of our 108 x 4 = 432 geometry), we illustrate the Fibonacci digital root sequencing for the 'Isolated Prime' candidates:
Analysis of the modulo 120 sequences discussed above reveals that the digital root of Fibonacci numbers indexed by squared members of natural numbers not divisible by 2, 3 or 5 always have a digital root of 1, while the squared numbers they're indexed by are always congruent to either 1(mod 120) or 49 (mod 120), i.e., The squares of all prime numbers >5 are congruent to {1 or 49} mod 120. This confines the distribution of squared prime numbers >5 to a sequence that is 2/120 or 1.66% of natural numbers. (The author has documented this sequence on the Online Encyclopedia of Integer Sequences [Sequence #A227863], where it's named: Numbers congruent to {1, 49} mod 120.)
This sequence is mysteriously prismatic, as prime numbers embedded in squares are split off in two angles (determined by the terminating digits of their factors), i.e., within a modulo 120 factorization wheel where 1 = 3°, the squares distribute to 3° and 147°, and are thus separated by 144° (And we have to ask whether it's coincidental that the digital roots of the Fibonacci numbers that form the 32-beat repeating palindromic sequence also sum to 144, not to mention, as we've already noted above, when the first three twin prime candidate pairs are summed and multiplied together, 24 x 36 x 60 = 51840, and 51840/360 = 144. This last fact will take on even greater significance when we diagram the 'square roots of squares,' second image, below. One more riff on 144: it's the 12th Fibonacci number following 89 and the only Fibonacci number that's a square other than Fibo1. When you divide 144 by 89, a number with magical properties we explore under 'Fibonacci Twins,' below, you produce an approximation of the golden ratio, or phi, with a difference of only +0.000056460660007.). Another fascinating feature of this array is that any even number of–not necessarily contiguous–factors drawn from any one of the 32 angles in this modulo 120 configuration distribute products to 1(mod 120) or 49 (mod 120), along with the squares. Two examples: 1) at 33°: 11 x 131 x 251 x 371 = 134187361, which is congruent to 1 (mod 120) ; ... 2) at 291°: 97 x 217 = 21049, which is congruent to 49 (mod 120). The matrix below illustrates these relationships:
Referencing the matrix above, the illustration below sieves the squares congruent to {1,49} mod 120 and takes their square roots. Twin prime candidates are emphasized to show their repeating zig-zag sequencing on the rungs of a perfectly counter-balanced lattice structure:
It is well known that the ratio of any Fibo(n+2)/Fibo(n) will converge to a limit φ + 1 = φ^{2} (an irrational number) as n approaches infinity (φ being the symbol for phi or the "golden ratio" which = 1.6180339887498948482045868343656 ...). And so it follows that the ratios of the Fibonacci numbers indexed to the twin primes (n and n+2) and/or twin prime candidates in sequence converge accordingly (in other words, the square roots of the ratios of Fibonacci numbers indexed to the twin primes and/or twin prime candidates in sequence converge to φ).
Here are some examples (square roots not taken):
Fibo(13)/Fibo(11) = 2.617977528089887640449438202247 ...Here's an interesting connection to Fibo(13)/Fibo(11) = 2.617977528089887640449438202247: Fibo(13) = 233, while Fibo(11) = 89. The square root of Fibo(13) divided by Fibo(11) = 12^{2} divided by 89 (and 89 has the basis of the Fibonacci sequence in the decimal expansion of its reciprocal). Then consider that 3 x 12^{2} = 432, a fascinating number we discuss above under 'Cirque de Primes,' below under 'Fibonacci Twins,' as well as on our page detailing the Magic Mirror Matrix where we also examine the amazing properties of number 89.
One other twin prime related deterministic algorithm worth noting stair-steps its way up the twin prime distribution channels with a periodicity of 6. We've dubbed these "perfect twins," given that the square root of their 'perfect' square sums is the first perfect number, 6, or one of its multiples. To calculate this interesting twin pair candidate sequence, start with x = 6 (then add 6 and repeat for each successive step, as shown below):
_________________ | |
√ | (x^{2}/2 - 1 + x^{2}/2 +1) |
x=6: 6^{2} = 36; 36/2 = 18; 18-1 = 17; 18+1 = 19; thus, twin pair candidates = 17 and 19; take the square root of 17+19 which = 6
add 6 (6+6 = 12)
x=12: 12^{2} = 144; 144/2 = 72; 72-1 = 71; 72+1 = 73; thus, twin pair candidates = 71 and 73; take the square root of 71+73 which = 12
add 6 (12+6 = 18) {repeat ... n}
The table below shows the first 20 steps in the sequence. All "perfect twins" are hi-lited in gold. [Note: The table is arrayed in five columns to show the +30 and repeating modulo 30 vertical incrementation. Also note that the digital roots for all prime pair candidates and/or prime pairs in this sequence equate to 8 + 1 = 9, e.g., 17, or dr(8) + 19, or dr(1) = 36 or dr(9)].
[Note: We've documented these "perfect twins" on the Online Encyclopedia of Integer Sequences (Sequence #A232878) where they're defined as Twin prime pairs which sum to perfect squares.]
Alluding to the Twin Prime Conjecture, given that all composite numbers extant in the infinite set of natural numbers not divisible by 2, 3 or 5 (which by definition contains all prime numbers ≥ 7 and their multiples and is defined as 1 {+6 +4 +2 +4 +2 +4 +6 +2} {repeat ... ∞}), can be produced algorithmically, employing eight geometrically-expanding factorization sequences, ultimately leaving the equal distribution of all prime numbers ≥ 7 ... n along eight diagonals; and given that six of these diagonals form the three twin prime distribution channels (described above), which, when combined into a single number line, reveal a divergent series: 11 {+2 +4 +2 +10 +2 +10} {repeat ... ∞}; and given that the infinity of prime numbers is proven, it's axiomatic that there is an infinity of prime pairs. Put differently, prime pairs, starting with [11, 13] are sequenced within a divergent (aka harmonic) series while paradoxically converging to φ^{2} when indexed to Fibonacci numbers like meshing gears.
Another interesting aspect of the three twin prime distribution channels becomes apparent when the eight radials of the Prime Spiral Sieve (or modulo 30 wheel factorization), are superimposed upon a {10/3} regular star polygon (see illustration, below). Thirty (30) is the total unit length of the ten straight line segments used to construct the {10/3} star polygon, where the straight lines connect every 3rd point of 10 equally spaced points lying on a circle's circumference. Note that three points of the polygon, 144°(12), 216°(18) and 360°(30), precisely and symmetrically split the twin prime distribution channels. And as you rotate this object, several other symmetries become apparent, e.g.: between 7 → 17; 13 → 23; and 7 → 23.
Even more interesting, is the result when you superimpose a 15-point star over a Prime Spiral Sieve with its 'aperture' increased to modulo 90. As shown below, every one of the 9 twin prime distribution channels is hit dead center by points of the star:
This superimposition reveals some interesting circularities: When you sum the 9 angles that hit the twin prime distribution channels dead center you get 1800. 1800/360=5; 1800/9=200. In turn, sum the 6 angles that hit equidistant between the twin prime distribution channels and you get 1080. 1080/360=3. 1080/6=180. In toto, 1800+1080=2880; 2880/360=8.
Before leaving the twin primes, let's take one more look at them from a digital root perspective. The illustration below shows the infinitely repeating digital root sequences for each of the 8 radii of the spiral sieve, as well the sums of the digital roots (in yellow) between each radius, including the twin prime distribution channels. It's notable that the twin pair digital sum sequences, {3,9,6}, {9,6,3} and {6,3,9} when positioned adjacent to each other in ascending order, create a "magic square" (in that all vertical, horizontal and principal diagonal sums equal 18), as pictured below:
The {3,6,9} triangulation is a member of the so-called 'Trinity of Triangles,' along with {1,4,7} and {2,5,8}. Looking at the graphic above, you'll note that each of the three Twin Prime Distribution Channels has one leg that sequences as {1,4,7} and the other as {2,5,8}, albeit the starting digits may vary, viz. {1,4,7} vs. {4,7,1}. When you combine the three legs that sequence as {1,4,7} as well as the three legs that sequence as {2,5,8}, these also form magic squares, as illustrated below. Another magic square hidden in plain sight becomes evident when you examine the sequences you encounter in rotation: {1,4,7} ... {2,5,8} ... {3,6,9} in various sequence combinations. Taking the digits 1 thru 9 these three triangles represent, they can be configured into a magic square representing 'et alia', in this case the famous 'Lo Shu Magic Square':
As discussed earlier, twin primes come in three digital root configurations: {2,4}, {5,7} and {8,1}. These dyads, in turn, can form the basis of 24-length repeating cycles given that (and there's a proof for this) any initial dyad (a(0),a(1)) of nonzero single-digit numbers enters a periodic cycle of length-24, except for eight pairs that enter cycles of length-8, {(3,3), (3,6), (3,9), (6,3), (6,6), (6,9), (9,3), (9,6)}, and one dyad, (9,9), which is immediately periodic of length-1. (Source: Derived from a comment posted by Jonathan Vos Post on the Online Encyclopedia of Integer Sequences, which the undersigned has edited, given seven of the eight length-8 cycles listed above were unwittingly omitted.): OEIS #A030132: Digital Root of Fibonacci(n)). For example, the first two terms of the Fibo digital root sequence just referenced, (1,1), enter a periodic cycle of length-24, as follows: {1, 1, 2, 3, 5, 8, 4, 3, 7, 1, 8, 9, 8, 8, 7, 6, 4, 1, 5, 6, 2, 8, 1, 9} {repeat}. So with these cycles in mind, take the three twin prime digital root dyads under discussion, sequence their 24-length digital root cycles, then tier and align them, and you'll discover a direct means to constructing the Trinity of Magic Squares via the Trinity of Triangles, as illustrated here:
[Note: Triangles and magic squares similar (if not identical) to those illustrated above can be derived from the digital root sequences of all three twin prime distribution channels. This is also true of dyads formed by paired radii of the Prime Spiral Sieve that sum to 30, i.e., numbers congruent to {1,29}, {7,23}, {11,19}, and {13,17} modulo 30 have variations on the 24-length cycling theme, with results similar to the above in terms of generating triangles and magic squares.]
And we'll end our exploration of twin primes with two powerful magic squares. First: The Trinity of Magic Squares x 3 can be constructed into a magic square where all columns, rows and principal diagonals equal 45 (and were you to construct an 8x8 square employing 64 Trinity Magic Squares shown below, all columns, rows and principal diagonals would equal 360).
Next we show how the Lo Shu Magic Square and Trinity of Magic Squares can be combined in a "4-Fold Way" to create their own magic square when configured in a 4 x 4 matrix of magic squares consisting of 144 elements where all rows, columns and principal diagonals sum to 60 and the sum total of the matrix's elements = 720 = 360 x 2 (and were we to construct a 6x6 square employing 36 4-Fold Way Magic Squares shown below, all columns, rows and principal diagonals would equal 360).