Bulletin of the American Iris Society, 145: 3-11 (1957)
Cytogenetics of Median Bearded Irises
L. F. R
ANDOLPH

Among approximately 400 crosses made last year by more than 20 hybridizers who are members of the newly-formed Median Iris Society, there were most of the possible combinations of the 8, 12, 16, 20 and 24-chromosome dwarfs with the diploid and tetraploid talls as well as various combinations of the dwarfs with the Oncos, Regelias and Arilbreds of mixed ancestry. Intercrosses among the different categories of dwarfs, and selfs and sib crosses of the rare species were also made to build up and improve breeding stocks.

In any attempt to evaluate crosses of this sort, most of which involve either interspecific combinations or hybrids of species and garden varieties, two major considerations are of paramount importance. The first of these concerns crossability in relation to chromosome number — whether viable seedlings can be obtained and if so will they be of value as garden varieties or useful for further breeding. The second consideration has to do with the manner in which specific characteristics are inherited. To predict what may happen in a given situation involving crosses of this sort, it is not enough to know about Mendelian ratios; one must also know about the pairing behavior of the chromosomes. With certain types of pairing there may be no Mendelian segregation whatever and, where differences in chromosome number are involved, ordinary dominance relations may be complicated by unusual dosage effects. These problems will be considered one at a time.

Chromosome Number and Breeding Behavior

From the standpoint of chromosome relationships and breeding potentialities the most promising crosses were those in which both parents have the same chromosome number. In a genus such as Iris where related species cross more readily than in most other plant genera, the general rule that those with the same chromosome numbers cross most readily and are most apt to produce fertile hybrids is definitely applicable. The one exception that might be said to prove the rule are those cases in which both parents are tetraploids. Then it doesn't matter very much whether they have the same or different chromosome numbers; their hybrids usually are partly or fully fertile because the chromosomes of the two sets contributed by each parent can pair together with a degree of regularity that assures gametes with balanced chromosome numbers.

From a casual glance at the lists of crosses submitted it appears that the significance of this rule about crossing things with similar chromo­some numbers is not widely appreciated, or iris hybridizers are unusually venturesome and willing to try anything once. Numerically, there were many more crosses attempted between parents with unlike chromosome numbers than with like numbers. Perhaps this shouldn't be deprecated too vigorously because in the past very significant results have come occasionally from such crosses, as for example the Parisiana by gatesii (24x20) cross that produced the famous William Mohr variety.

There were literally dozens of pumila/tall crosses made last year and when the results of these are added to those of previous years, we will certainly have a clearer picture of the range in color forms and growth characteristics to be expected among their hybrids. Extreme types among the talls, such as Black Forest and Snow Flurry, when crossed with the deep violet pumilas should produce very interesting seedlings since both are known to be outstanding parents.

Crosses of pumila with recessive types of tetraploid talls, such as the plicatas, tangerine, pinks and whites from blues and plicatas have already disclosed the very interesting fact that the chromosomes of this dwarf species are carrying as hidden recessives some of the same genes that are to be found in the talls. Having checked the chromosome count of one of the plicatas from the rather sensational Mariposa Mia x Cretica cross it is definitely known to be a true 40-chromosome hybrid.

A word of caution with respect to the use of Cretica in breeding might not be out of order since there are mixed reports concerning its value as a parent. This pumila derivative has a certain charm of its own as a garden variety, and this may account for the fact that it was found growing on the island of Crete far south of the known range of the wild pumila, with which its chromosomes are identical. Perhaps it was taken to Crete by an iris-loving emigrant from the Balkans and later escaped from cultivation and became naturalized, as other iris species have done elsewhere.

It is well known that most of the Lilliput Hybrids thus far obtained more closely resemble the dwarf parent than they do the tall parent, with respect to the height of the bloomstalk; but unfortunately, their foliage is usually not as definitely dwarfed. It is not so well known that the pumila/tall crosses are capable of producing miniature dwarfs more tiny than any known derivatives of pumila itself. These tiny hybrids form dense grass-like tufts of foliage no more than two or three inches in height. Although apparently healthy and vigorous they rarely produce bloom stalks according to Bee Warburton from whom specimens were received for cytological study. Their hybrid status has been verified by chromosome counts which have shown them to have the expected 16 chromosomes from the pumila parent and 24 from the tall parent. It is well known that wide crosses may sometimes produce atypical hybrids of this sort.

The crosses which have involved the true 48-chromosome aphylla and the tetraploid talls should be very interesting since both contribute two sets of 12 chromosomes to their hybrids. No difficulty is involved in making such a cross but rather poor seed germination has been reported from some of them. However, this species is now available in this country from at least three widely separated geographical areas in the Balkan and Middle East, and some combinations will undoubtedly produce better results than others.

The unique type of branching extending to the base of the 6 to 10-inch bloom stalks of aphylla makes this species very different in appearance from the unbranched 40 and 48-chromosome dwarfs. Some indication of the amount of branching to be expected from the hybrids of aphylla and other 48-chromosome dwarfs can be obtained from what is known concerning the wild hybrids of aphylla and pumila reported from central Rumania where the ranges of these two species are sympatric (found in the same localities). An intermediate amount of branching sufficient to display the terminal blooms to better advantage is characteristic of these hybrids.

A whole new series of fertile hybrids can be confidently expected from combining the tetraploid aphylla with both the 32 and 48-chromosome tetraploid dwarfs. In fact they are already on the way.

At the diploid level there appears to be little difficulty in obtaining seed from intercrosses of diploid dwarfs such as attica and pseudopumila (n=8), or bosniaca, reichenbachia, mellita, and rubromarginata (n=12), These dwarfs also cross readily with diploid talls and data from advanced generation hybrids indicate that the F1 seedlings of certain combinations (mellita x 2n talls) are sufficiently fertile and there is enough pairing of their chromosomes to provide the necessary mechanism for recombinations of traits of garden value.

The diploid dwarfs apparently do not cross any more readily with the diploid arenaria than do the tetraploids or the 40-chromosome dwarfs. The attica/arenaria hybrid reported by Darby in the 1954 DIS Portfolio adds a new chromosome number (2n = 19) to the ever-increasing list of new kinds of hybrids, as does the Pumar hybrid (pumila/arenaria) reported by Ackerman in the 1956 DIS Portfolio. Cytological examination of this hybrid in my laboratory showed that it not only had the expected chromosome number (2n=27) but in addition, the positive identification of certain arenaria and pumila chromosomes was possible. Add to these new hybrids of arenaria and the miniature dwarfs those produced years ago from crosses with the 40-chromosome dwarfs (Mistopink, Tiny Treasure, Keepsake) and it becomes obvious that, although arenaria is one of the few species in related sections that does not cross readily with the dwarf bearded irises, repeated trials are worthwhile.

Genetic Analyses

The simplest cases of inheritance involving individual gene difference are those in which recessive characters make their appearance among the seedlings of parents, one of which displays the character in question and the other does not. Such cases include an element of surprise, especially when a wide cross is involved and there is no particular reason to suspect that the non-recessive parent may be heterozygous for the character.

Currently, the plicatas that appeared among seedlings of Cretica crossed with tetraploid tall bearded plicatas have created much discussion, and the opinion has been expressed that these could not have been true crosses. But as previously stated, the chromosome count has established that it was a true cross in the case of Earl Roberts' plicata seedling from Mariposa Mia x Cretica; also, the plicata, Dale Dennis, is reported by Dorothy Dennis, the originator, to have had a similar origin. Apparently, Cretica is heterozygous for the same plicata gene for which Mariposa Mia is homozygous.

There have been consistent reports from several hybridizers that there is little or no segregation for height differences in the first backcrosses of pumila/tall hybrids to the tall parent, or in F2 progenies from sib crosses of the first generation hybrids, and my own observations confirm these reports. This is to be expected if the 16 pumila chromosomes and the 24 chromosomes of the tall parent present in the first generation hybrid regularly pair among themselves (autosynapsis), and the members of each pair are not heterozygous for genes controlling height differences.

The cytological observations of Randolph and Heinig reported in AIS Bulletin 118 have in fact shown that the 40 chromosomes of the pumila/tall hybrids usually do form 20 pairs. If allosynapsis occurred regularly there would be 16 pairs and 8 univalents, i.e., the 16 pumila chromosomes would pair with 16 of the 24 chromosomes of the tall parent and the remaining 8 chromosomes of the tall parent would have no partners with which to pair.

In the second backcross of the pumila/tall hybrids to talls the amount of segregation for height differences would depend on the manner in which the 8 pumila chromosomes of the first backcross plants were distributed to the gametes and the potency of their genes. It is probable that difficulty would be experienced in obtaining many vigorous seedlings from such crosses, if the 44-chromosome, first backcross plants are no more fertile than the comparable 44-chromosome hybrids of the 40-chromosome dwarfs and talls. However, it is reported that second backcross progenies have been produced and have included late-blooming dwarfs. Miniature dwarfs also have been reported from backcrosses to pumila.

Reports that tangerine pinks are appearing in backcrosses of pumila/tall hybrids to tall pinks implies that pumila is carrying these same genes for pink, if autosynapsis is the rule in these hybrids. Geddes Douglas is said to have a pink from the Lilliput Hybrid, "Wee One" x Pink Formal ready for introduction. Assuming that autosynapsis occurs regularly in the pumila/tall hybrids, one of the pumila chromosomes and one of the tall chromosomes of the hybrid would have to be carrying a recessive allele of pink in order to obtain pink seedlings in backcrosses to a pink. If pumila does not carry any recessive genes for pink and if the tall parent of the pumila/tall hybrid was a pink, the appearance of pinks in the backcross progeny would indicate that allosynapsis had occurred. This is an interesting example of the manner in which chromosome pairing can affect Mendelian segregation.

A well documented case involving unusual dominance relations of genes for anthocyanin color in pumila/tall hybrids was described by Paul Cook in the 1952 DIS Portfolio. Since additional examples of this very interesting type of genetic effect may be expected in other iris species crosses, the principles involved deserve special attention. A cross of a blue tetraploid tall bearded seedling No. 10942 by a blue pumila seedling produced 19 blues and 24 non-blues (yellows, creams and whites), and from this same 10942 tall seedling pollinated by a yellow non-blue pumita seedling he obtained 50 non-blue seedlings. From these results Paul concluded that both the blue and yellow pumila carried a gene which inhibited the blue of the tall seedling but not the blue of pumila. The possibility of a recessive gene for white was excluded because a 1:1 ratio rather than a 3:1 ratio of blues and whites was obtained from the intercrosses of the two blues and there were all whites rather than a 1:1 ratio of blues and whites from the other cross. A dominant inhibitor of pumila and tall anthocyanin carried by the yellow pumila in the homozygous condition would explain the absence of blues in its progeny but could not have been involved in the other cross, both parents of which were blue.

Additional data from test crosses which substantiate his interpretation were supplied recently by Mr. Cook. When the non-blue seedling No. 1249, which came from 10942 x blue pumila, was crossed with Green Spot (10942 x yellow pumila) there were whites, creams and yellows but no blues among the 20 seedlings. Green Spot x Shining Waters, a tall blue, produced 8 non-blue and 5 pale blue seedlings. The tall blue Distance x 5148, a non-blue from 10942 x blue pumila, produced 27 non-blues and 4 pale blues. Although no additional data are available from back crosses to the blue pumila parent, those from the back crosses to talls support Mr. Cook's interpretation.

The startling appearance in tall/pumila hybrids of green colors much brighter than any which have been obtained in the talls suggests that pumila has a unique genotype capable of modifying blues in the presence of yellow to produce greens like that of Green Spot, which resulted from the blue tall 10942 x yellow pumila cross mentioned above. Also, the red-purple spot or signal patch often seen in the center of the falls of yellow pumilas is a character which may persist in crosses with the talls, but has not been reported in seedlings of talls.

Since the pumila blues appear to be qualitatively different from most of the tall blues, the view is prevalent that they are genetically different. However, the examples just cited of other color differences suggest that in different genetic environments, such as are found in pumila and the tetraploid talls, the same genes may exhibit different phenotypic effects. This is known to happen in other plants. In fact, one species of cotton has a conspicuous signal spot not unlike that of pumila and various Oncos and Regelias, which cannot be recovered when introduced into another species.

Genetics of White Irises

A pure white color form of I. attica collected on Mount Parnes near Athens, Greece. This species typically has no stem, the blooms being raised several inches above the falcate leaves by an elongate perianth tube enclosed by elongate spathe valves.

Because of the importance of white irises in breeding programs it is essential to know whether one is dealing with a dominant or recessive white. For hybridizers interested in developing new kinds of Median irises it is especially important to know whether the same white alleles, either dominant or recessive, occur in the dwarfs and talls. In each of the three classes of dwarfs, the 16-chromosome attica and pseudopumila, the 32-chromosome pumila and the various 40-chromosome derivatives, there are white irises, but not very much is known about their genetic behavior. Among the numerous color variants of attica collected on Mount Parnes during my visit to Greece in the spring of 1954, there was a nice white, but in my garden it is not as vigorous as most of the other clones that were collected from the wild populations near the top of the mountain. Crosses have been made with yellows, blues, and violet-purples to find out if this white attica is a dominant or recessive.

In pumila crosses Walter Welch obtained a white from two blues which is probably a recessive (c.f. Portfolio 5: 18, 1954) and in the 40-chromosome dwarfs he reports all whites from selfing Bouquet, and a similar result from crossing the white Bouquet with a white seedling of Fiancee and Fairy, two other whites. When describing these results in the 1954 Portfolio, Walter stated that he had never found any indication of a dominant factor for white in the chamaeiris dwarfs. However, in this same article he reported that Harbor Lights selfed gave 11 yellows and one purple, and from Excelsa selfed he obtained 11 yellows and 3 purples. The appearance of purple seedlings in the selfed progeny of non-purples with these frequencies would be difficult to explain on any other basis than that a dominant inhibitor of anthocyanin was involved. The fact that these varieties are yellow rather than white is irrelevant since the yellow color is controlled by a different gene.

Evidence that this same type of anthocyanin inhibitor is present in the pumila dwarfs is available in the variety Blarney introduced by Welch several years ago; this variety has appreciable amounts of anthocyanin tinged with green. One parent of Blarney was Cook 1546, a cream colored amoena of pumila derivation, and the other parent was a white seedling of Fiancee x Fairy, believed by Welch to be a recessive white. None of this evidence of the existence of a dominant white inhibitor of anthocyanin in these dwarfs in any way controverts Dr. A. H. Sturtevant's well-known theory that the dominant whites of the tall bearded irises came from the dwarfs; nor is it adequate proof of the correctness of his theory.

At least one recessive white allele of colored anthocyanin found in diploid tall varieties, Solitaire and White Knight is also present in the 40-chromosome dwarfs. Looking back through my crossing records of the past several years, I find that white seedlings appeared in a family of Morocain x Solitaire seedlings and crosses of Solitaire and White Knight produced all white seedlings.

The genetic relationships of the various dominant and recessive whites in the dwarfs and talls can be worked out most easily at the diploid level and then applied to the more complex polyploids. With this in mind while collecting European wild irises in the spring of 1954, it was especially interesting to find, in a population of I. pallida not far from Dubrovnik in southern Yugoslavia, a pure white plant among otherwise typical lavender and rosy lavender clones of this species. The circumstances under which it was found suggested that this white pallida was a mutant of recent origin, which could be the same or different from the recessive whites among cultivated diploid tall varieties having pallida ancestry.

The relation of yellow color to white, blue, purple and the various blends which add so much variety to iris colors, has interesting genetic implications. It has been known almost from the very beginning of hybridizing activity with irises, that the blends are obvious mixtures of yellow and anthocyanin. In the tall bearded irises the purples and reds segregate yellows or creams and varying shades of lavender and blue, but in the pumila dwarfs the purples seem to behave differently — as if they are not mixtures of blue and yellow.

The clear, pure yellows exist only in the complete absence of anthocyanin color, that is, in combination with white, which may be due to any one of several recessive alleles of anthocyanin or to a dominant inhibitor of anthocyanin color. Yellow itself ordinarily behaves as a dominant and the recessive is white, phenotypically indistinguishable from other whites.

However, there is some evidence in the tetraploid talls that yellow can result from intercrosses of whites. This suggests complementary gene action, by which is meant that more than one gene is involved, any one of which is incapable of producing yellow color but when combined in the same genotype, yellow color is produced. The additional possibility that yellow may be a recessive is suggested by the appearance of yellow seedlings in a selfed progeny of April Morn, a blue pumila dwarf. If this is substantiated, it will add another unusual type of genetic behavior to a growing list already attributed to this very interesting dwarf species.

From the foregoing comments on the cytogenetics of dwarf, intermediate, and tall bearded irises it should be obvious that there are many unsolved problems to challenge the ingenuity of hybridizers interested in creating new kinds and delving into the hidden secrets of long-time favorites.