Can't be certain off the top of my head, but is this not another case of a virus causing changed floral colouration?

not coincidence or goot grafting. Could be viral but tend to think variation of the hybrid cultivar "Kaliedescope". Some are not very stable.

I don't think viruses are involved in this variegation.

It is likely to be based on transposons and therefore the colour depends on which pigment genes the transposon jumps into (turning it off) or jumps out of (turning it on).

Did no one take biology in junior high school? The genetics of 4 o'clocks is one of the basic teaching tools. Mirabilis jalapa exhibits what is called incomplete or co-dominance. Unlike other flowers that exhibit co-dominace, Cleome or Snapdragons, which say when you cross a red homozygous and white homozygous plant you get pink. 4 o'clocks exhibit codominace in floral variegation which is rather spotty in expression. Some of the flowers may be half one color and half the other, you may have striped flowers, you may have solid colored flowers of both colors and variegated flowers on the same plant, you may even have mixing of colors with variegation (pink flowers with red and white stripes).

It's not a virus.

floral variegation which is rather spotty in expression. Some of the flowers may be half one color and half the other, you may have striped flowers, you may have solid colored flowers of both colors and variegated flowers on the same plant, you may even have mixing of colors with variegation (pink flowers with red and white stripes).

Floral variegation is a typical result of transposons. The problem with variegation in flower pigment is that all cells have the same genes for the pigment (except in chimeras). Therefore normal/classical Mendelian genetics cannot explain the variegation. Dominance/codominance apply to all cells equally.

One test of whether the variegation is due to a transposon is to examine the pattern of variegation in terms of the growth of the flower during development.

Most flowers develop from a meristem at the base of the petal which divides producing a row of cells. These cells grow and elongate. At each mitotic division a transposon has a chance to be cut-out of the location it is currently in. Often a pigment colour mutation is due to the presence of a transposon in the pigment gene and turning it mutant and ineffective. When the transposon is removed during mitosis (jumps out) the gene becomes normal and then that cell begins to produce pigment. All the descendents of that cell also produce pigment. If a variegated petal shows some streaks of pigment that are small/narrow towards the petal base and widen as they extend to the petal edge that is a strong indication that a transposon was removed in the cell at the very base of the streak and all its descendents produced pigment.

In a red flower x white flower cross if all the offspring show red flowers then red is dominant. If all the offspring show pink flowers then red and white are codominant. Mirabilis jalapa shows this type of codominance.

Here is a quote, from http://www.biologie.uni-hamburg.de/b-online/e08/08a.htm,
"Incomplete dominance in flowers of Mirabilis jalapa AA genotypes have red, Aa genotypes pink and aa genotypes whitish flowers (redrawn from C. CORRENS, 1902)"

The web-site includes a diagram showing solid pink flowers without variegation. Incomplete dominance or codominance cannot explain variegated individual flowers. Transposons, flower-breaking viruses and ploidy differences in cell layers and possibly certain types of nuclear genes interacting with mitochondrial/chloroplast genes can explain flower variegation.

The link below is a review of plant variegation and transposons and briefly mentions the flower variegation in Mirabilis jalapa.

Here is a link that might be useful: PDF review of transposons - Mirabilis