56 PNL Volume 17 198 5 RESEARCH REPORTS
Marx, G. A. NYS Agricultural Experiment Station, Geneva, NY USA
I have long been captivated by two early, landmark papers by Hans
and Olof Tedin which describe the remarkable multiple effects of the
flower color gene ar (1,2). Because the effects produced by ar are
similar in some respects to those conferred by am-1 and am-2, two genes
which have also occupied my attention for some time, I decided in 1979 to
analyze anew the abstruse and artful antics of ar.
The flowers of ar plants have a distinct blue hue in contrast with
the purple violet flowers of the wild type. Like several other flower
color mutants, ar exerts a pleiotropic effect on axil color (also blue).
Mutant ar also interferes with the expression of oh, a seed gene which
normally produces self-colored reddish brown seed if A is also present.
Oddly, ar in combination with b (pink flowers and axils) results in very
pale pink flowers (in my hands white or nearly white whether cultivated
in the field or glasshouse). Thus, neither ar nor b is epistatic to the
other with respect to flower color; instead, the two genes interact to
yield nearly colorless flowers. Though flower color is quenched in ar b
plants, axil color is only slightly less pink than in plants carrying b
alone. A more noteworthy feature of ar, however, is its pleiotropic
effect on seed structure. Specifically, the hilum of seeds from ar
plants is reduced to a narrow slit and the tracheid bundle (or bar) dis-
appears. But with the introduction of the recessive seed gene, z, the
hilum is restored to normal, without, incidentally, influencing flower or
axil color. This suite of effects, involving site specificity, pleio-
tropy and interaction, provides a potentially powerful model for studies
of gene regulation and expression. Moreover, as indicated above, the
parallels between the ar b z system and the am-1, am-2, b system led me
to acquaint myself with the former and, while gaining firsthand ex-
perience, to develop a series of desirable, genetically defined lines for
use in further work.
WL 251/, the type line for ar, was used as a parent in the initial
crosses made in 1979. Unfortunate1y , WL-25 is not only tall and late,
but also shows poor pod and seed set . Although other more "friendly" ar
lines were available in my collection, I wished to remove any doubt that
the starting material was indeed ar. WL-25 also carries oh, the
expression of which, due to the presence of ar, is partially inhibited.
WL-25 was crossed with, my line C879-348 (le A Ar oh Cry st La B). The F1's
and subsequent progenies from these crosses were characterized by ex-
tensive semi-steri1ity or complete sterility. Because of this, no
meaningful data could he collected and therefore no details will be
given, except to say that a progeny test oi one F2 segregant yielded an F3
progeny containing eight St plants and 73 st plants. Rigorous selection
was practiced but it was not until the F7 that a fully fertile, dwarf
line carrying ar, oh, and B was isolated. Even so, the seeds in this
line exhibited somewhat impaired development presumably owing to the
effect of ar on the hilum structure of the seeds. Meanwhile, however,
descendants from the original cross, starting with the F3, were used in
new crosses with a series of other lines. Certain segregants from these
second-order crosses were used in turn as parents in third-order crosses,
so to date some 70 crosses and over 700 progenies have been grown and
1/ Seeds kindly supplied by Dr. Blixt.
PNL Volume 17 1985
My observations agree in nearly all respects with those of H. and 0.
Tedin, but some details remain to be settled. A modest addition to their
findings can be mentioned: in addition to z, genes a and b also restore
normal hilum structure in ar plants. I have also introduced wa, a marker
for genes ar and oh, and n and was, markers for z, into the system.
Importantly, the material now at hand is "tamed", being represented by
fertile dwarf lines which are easy to cultivate and in which the gene
expressions are clear. A number of the lines are near isogenic for one
or more relevant genes and thus are particularly favorable for demon-
strating gene action and interaction. Finally, the material is well
suited as parents for new crosses with genes in the am-1 and am-2 system,
a project which is now underway.
1. Tedin, H. 1920. Hereditas 1:68-97.
2. Tedin, Hans and Tedin, Olof. 1928. Hereditas 11:1-62.