Purple Haze x P20 Project

from "Return of the Killer Tomatoes"


P20 is a line developed by Dr. Jim Myers and his graduate students at Oregon State University that segregates for purple fruit. It combines 3 (possibly 4) key genes to produce the "purple" color in the plants with the Aft gene being primarily responsible for purple fruit color.
Boches, P.S. and Myers, J.R.. TGC REPORT 57, 2007 pg 14
The OSU high anthocyanin lines were created by crossing the accessions LA3734A (containing vio, a radiation induced mutant with high anthocyanin stems and veins), LA3736 (containing atv, a gene introgressed from S. cheesmanii LA0434 that causes high anthocyanin in stems and leaves), LA1996 (containing Aft, a gene introgressed from S. chilense that causes anthocyanin accumulation in the fruit), and LA3668 (containing Abg, an anthocyanin fruit gene introgressed from S. lycopersicoides). Although most of our stable high anthocyanin lines (e.g. P20) have LA3668 (Abg) in their pedigree, they probably do not contain Abg, since Abg can not be maintained in a homozygous state due to an inversion on the lower arm of chromosome 10 in S. lycopersicoides relative to S. lycopersicum
The genes were incorporated from other tomato species using traditional breeding techniques only.

For more about that program, the genes involved and how it relates to other varieties like 'Cherokee Purple' and 'Purple Smudge' see
Purple Tomato FAQ (a pdf file).

For more about P20 and some of the other lines from that program see
Occurrence of Anthocyanin in Cultivated Tomato

The work at OSU is different than the work at John Innes Centre in the UK which produced another kind of purple fruited tomato via genetic modification of genes from the snapdragon flower.
Scientists Developing Cancer-Fighting Purple Tomatoes

2012 Update:

OSU has now released the line 'Indigo Rose'.
OSU Extension Release Notice

'Indigo Rose'

Israeli horticulturists release 'Black Galaxy'.
Link from DailyMail.UK 2-8-2012
"A dark day for the salad counter � Israeli scientists breed black tomato"

'Black Galaxy'


My Work:


I received P20 seed from Darrel Jones of selectedplants.com sometime before the spring of 2008. Darrel sells plants of some of my material and I have made crosses for him in the past. Darrel received seed from Dr. James Myers and was given permission to use P20 for breeding purposes. I am grateful to Darrel and Dr. Myers for sharing seed.

I began by segregating out P20 lines which expressed the most coloration at the seedling stage. This increased the likelihood the recessive atv gene would be present. I did so by subjecting the seedlings to cooler temperatures than I would normally grow my other tomatoes and high light conditions. These are the lines that got set out and grown for pollen donors. I made multiple crosses from those plants. Once fruit on the pollen parents colored up, I noted the more intensely colored lines (assuming these expressed the Aft gene). Crosses made with these pollen parents on set crossed fruit were selected for seed and I discarded the other crosses from the weaker pollen donor types (probably lacked the Aft gene). I took this effort because even at the seedling stage it was apparent that the P20 seed I received still segregated to some degree. So it does not surprise me that some have seen different results in taste, habit and coloration with P20. If one reads the link above to the study it notes how they also did some selection for an open habit type to let more light into developing fruit.

I made crosses to several lines that year. However I focused on crossing a beefsteak like selection of an F4 line derived from the F1 'Purple Haze' to a P20 segregated line. This is because that line had several traits I wanted to get with anthocyanin: great flavor, higher brix, gf green fleshed gene (which makes for purple/black/brown flesh), clear skin, potato leafed foliage and larger size. This was the only line from all the crosses I sent to the southern hemisphere to be advanced and get 2 seasons in one year. Here are some of the prelimirary results of that cross:

Here is what the foliage type I selected looked like after the cold treatment.
Plant on upper right (compared to "normal" and variegated potato leaf).

This is the primary pollen parent I used.
Notice it lost its purple foliage coloration once the season progressed.

click picture for photo of mature fruit

Here is an example of the Purple Haze type line used as the female parent

Here is a crossed fruit used to get seed (several lines were used)

The F1 seeds of P20 x "Purple Haze' F4 were grown in both South Africa and New Zealand (2008-2009).

Here is an example of F1 fruit from New Zealand.

Here are examples of two different lines from South Africa

incorrectly named 'Brandokee'



At this stage I grew out 200 seedlings and tried to select as I did in 2008 for the more intensely colored foliage.

A close up of cotyledons shows that there are some differences early on even at the cotyledon stage. This examples of cotyledons show 2 of the 3 variations. The more intensely colored form (left and slightly less intense now due to leaf shading) and intermediate (right - purple stripe in the midrib and/or purpling of the stem) The 3rd other form not shown is just all green. The degree to which they express at this stage is effected by temperature and how much light/shading the cotyledons receive.

Here are examples taken a few days later than the one above.
The regular leaf plant corresponds to the cotyledons on the left above.
The potato leaf example below is not the same as the one above.

click picture for larger photo

click picture for larger photo

About the only other thing worth mentioning about the foliage is the hairiness and the tendency of the base of trichomes to produce "purple spots"
(which can also been seen at times in "normal" seedlings if they don't get enough phosphorus).

At that point in 2009 I focused in on selecting for the atv gene to get nice purple coloration in the foliage. I grew out several trays of seedlings, subjected them to cool and high light growing conditions and got about about 10 good purple seedlings from 200. Some of those lines were crossed later in the season to incorporate other traits. Later that year I focused on the green flesh gene (gf which is what makes "purple/black" fruit in lines like Cherokee Purple - see purple fruit FAQ link above), clear skin, a potato leaf variant and decent size and/or multilocule fruit. If you do the math on the probabilities (excluding fruit size) the chances of finding what I want from the 6 different genes are about 1 in 4096!

I was hoping to get lucky.

from "Return of the Killer Tomatoes"

August 2009 Updates

I selected about 80 plants which all showed good purple color in the foliage and planted them at various locations. The amount of anthocyanin seems to vary in about 4 different classes ranging from none at all, slight, moderate and heavy color. Only about 10% of all the lines produced a clear epidermis. Of those only two produced heavy or moderate "purple" and potato leafed plants. The fruits with yellow skin and the green fleshed gene produced some fruits that have a muddy brown color to them. Some of these have a very good flavor to them so I will select a few to grow for next year.

The taste varies considerably. Some are very tart and have a slight aftertaste. Some are sweet and have a very good flavor. Several brixed (a measure of sugars) greater than 7%.

There were a few that produced beefstake type fruit.

Here are some of the final pictures of various fruits.

Purple Haze F1 line (foreground) vs P20PHPL18 line (background)

P20PHPL18 interior

clear skinned lines with "purple"

#3 is clear skinned and produced no "purple" coloration
This grouping is an example of the 4 degrees of coloration

Above's Interiors

An example of the beefsteak type (multicarpellate)

And here is a beautiful unripe beefsteak that stood out from good sun exposure

Additional 2009 Work

In addition to growing out F2 material from the 'Purple Haze' x P20 lines I also grew out the F1 lines from last years crosses that did not get sent to the southern hemisphere.
Some of the F1 lines were:

For size:

For earliness + sweet

For sweet + gf trait

For variegated foliage:

For bi-color flesh:



Yellow and green fleshed fruits with purple segregated in this year:

Yellow flesh + clear skin line

This green fleshed line was more shaded so the anthocyanin expression wasn't as intense when I took this photo.

This is a Yellow/Gold (yellow flesh + yellow skin) line:

A few different lines from one field.
Cracking was a problem after rain in this field with no irrigation.

Early in the year I grew out about 800 seedlings looking for the atv + alb genes (purple + variegated foliage).
Out of 800, I was able to select about 20. Here are some examples of what those leaves looked like:

Purple potato leaf seedling without variegation.

Purple potato leaf seedling with variegation.

Purple potato leaf with variegation.

60x magnification of anthocyanin leaf surface

Crossing in 2010

Several P20 advanced lines were crossed to high brix Asian lines and a F1 Anthocyanin producing cherry line was made.

Here is the F1 anthocyanin producing hybrid in 2011.



2011 was a tough year. My home plot, which held the majority of this project near a fence line, was almost a total loss due to deer and groundhog damage. A had few backup lines at another farm survive. The yellow line produced amazing flavor and did not have as much cracking/bursting. It is stable enough that I named it and distributed some seed. It is named 'Manduh'.

e mau ke aloha

The other line was the F1 hybrid: high brix, clear skinned cherry. Several sister lines cracked/burst but one line did well with little cracking. It had "sister lines" because one of the parents was not stable. This farm is a tough site because it has no irrigation. It was also a tough year with extreme temperatures and little to no rain for the month prior to ripening. So when it did rain many lines burst.

These fruit developed in the shade so did not color up well.

Sister line: Note bursting of skin.



2012 was a tough year for deer damage. In 2013 a dvancement was made on stabilizing several lines noted above and further selection on earlier generations and salvaged lines from 2012 has produced a very nice looking, tasty and sizeable beefstake line.


Just 4 key advanced families (size and greenfleshed background) were selected and grown out on a large scale looking for the best flavor and flavor/yield combinations. This was a difficult year for flavor as this summer was unusually cool and rain came right at key ripening. Hopefully the better tasting lines in a poor year trasnlates to better in normal seasons.

At home advanced brix lines were backcrossed to high brix parent lines or with lines from diverse high brix parents combined with anthocyanin expression to furhter develop flavor and work out some of the genetic drag coming from the source Aft or atv parents.

Thanks to
Dr Myers, OSU
David Hallett in NZ
Geoff Dunn in South Africa
Darrel Jones selectedplants.com
Kurlbaum's Heirloom Tomatoes

back to the tomato page