Jour. Am. Soc. Hort. Sci. 38: 634-654 (1940)
Crossing Relations of Some Diploid and Polyploid Species of Roses

By J. C. RATSEK, W. S. FLORY, JR., and S. H. YARNELL,
Texas Agricultural Experiment Station, A. and M.
College of Texas, College Station, Tex.

ROSA is one of the plant genera in which many polyploid species are known to occur. In the Texas Experiment Station's breeding program to secure improved rootstocks for roses a rather large number of interspecific crosses have been made or attempted. The relations apparent when the crosses among the diploid species alone were considered have been reported earlier (7). The present paper takes up the crossing relationships of species where forms observed or reported to be polyploid are involved. In these crosses the materials and methods have been identical in source and use as described for the studies among diploids alone (l.c.).

A considerable number of seedlings, of sufficient size for character study, have become available from the earlier crosses. Scrutiny of this F1 material indicates that the great majority of the seedlings from the controlled crosses are true hybrids.

Where there is no evidence to the contrary we have considered the relative ease with which a cross is effected as indicating the degree of genetic affinity and relationship existing. Where the Caninae roses are seed parents, however, it is necessary, despite our seedling observations, to consider the possibility of the frequent heavy seed setting as being due to some type of apomixis and not to actual fertilization (6, 9, 10). The data are presented so as to show the compatibilities existing between different species, or between different groups of species. The data from the crosses have been considered when the parent species were grouped from each of the following three standpoints: (a) chromosome numbers;- (b) taxonomic sections (following Rehder's classification) and; (c) genetic groups (following Hurst's system).

RESULTS AND DISCUSSION

1. Parental Species Grouped According to Chromosome Numbers

For reported chromosome counts reference has been made to the papers of Blackburn (1), Erlanson (2), Hurst (4, 5, 6), Tackholm (9, 10), and to the chromosome lists of Tischler (12), and Gaiser (3). In many cases, and especially in instances of questionable material, chromosome counts have been made on our parental forms. Where there was good reason to doubt the authenticity of named material or of reported chromosome numbers of either parent of a cross, the cross is excluded from present consideration unless verification has been possible.

*CybeRose note: This is probably Prof. Allard's
OP seedling
from 'Harison's Yellow' (BBCD).

While report of the cytological findings will appear later we may note several points here which make for clarity in the tables. Among our three plants of Rosa blanda were two diploids (2n = 14) and one tetraploid (2n = 28). The tetraploid is not readily discernible from the diploids, and before the chromosome differences were noted some pollen mixtures of the three plants were used; data from such crosses were discarded for the purposes of this paper. In several cases chromosome differences were noted between species and their varietal forms, or between different varieties of the same species. As examples we may give the somatic chromosome numbers of the following: Rosa lucida, 28; R. lucida alba, 14; R. moschata alba, 14; R. moschata floribunda, 28; R. recondita, 14; R. recondita villosa, 28; and R. xanthina, 14; R. xanthina allard, 28.* Such counts as these have been taken into consideration in preparing the tables.

Abreviations used in Tables I to X inclusive are as follows:

R. coriifolia Froebeli = cor. Froe.
R. damascena trigentipetala = dam. trig.
R. foetida bicolor = foet. bicol.
R. gallica conditorium = gall. cond.
R. gallica splendens = gall. spl.
R. hibernica gravesii = hibernica grav.
R. moschata floribunda = moschata flor.
R. multifora cathayensis = mult. cathay.
R. multiflora chenault = mult. chenault
R. multiflora platyphylla = mult. plat.
R. multiflora upright = mult. upright
R. multiflora Welch = mult. Welch
R. spinosissima = sponosiss.
R. villosa recondita = villosa recond.
R. Wichuraiana (r'd) = Wichuraiana

TABLE I—CROSSES BETWEEN DIPLOID AND TETRAPLOID ROSE SPECIES

Cross
No.
Seed Parent
(Diploid)
Pollen Parent
(Tetraploid)
Flowers
No.
Crossed
Per
Cent
Set
Seeds Per Hip Per
Cent
Filled
Crossed O.P.
-39 Beggeriana (5)†DD‡ foet.bicol. (1) BBDD 11 0 — — —
-39 Beggeriana (5)†DD‡ xanthina allard (1) 14 0 —    
33-37 blanda B (5) DD foet. bicol. (1) BBDD 39 8 5.7 22.0 26
52-37 blanda B (5) DD gall. spl.** (2) AACC 22 62 22.4 22.0 102
53-37 blanda B (5) DD saturata* (5) CCDD 45 24 27.5 22.0 125
-37 Giraldii (5) EE foet. bicol. (1) BBDD 26 I — — —
-37 Giraldii (5) EE gall. spl.** (2) AACC 27 1 — — —
-37 Giraldii (5) EE oxyodon DDEE 23 0 — — —
-37 Giraldii (5) BE saturata (5) CCDD 47 2 — — —
-37 Giraldii (5) EE saturata (5) CCDD 25 0 — — —
71-38 Helenae* (6) AA foet. bicol. (1) BBDD 9 22 1.0 10.4 10
-39 Helenae* (6) AA foet. bicol. (1) BBDD 29 — — — —
60-38 Helenae* (6) AA Manettii (7) 10 10 1.0 10.4 10
77-39 Helenae* (6) AA xanthina allard (1) 13 15 1.0 10.4 10
20-37 lucida alba (4) bella* (5) BBEE 20 40 17.1 47.2 36
46-38 lucida alba (4) dam. trig.* (2) AACC 13 8 2.0 47.2 4
-38 lucida alba (4) foet. bicol. (1) BBDD 15 0 — — —
39-39 lucida alba (4) foet. bicol. (1) BBDD 12 25 4.7 47.2 10
18-37 lucida alba (4) gallica (2) AACC 23 4 2.0 47.2 4
-38 lucida alba (4) gall. spl.** (2) AACC 13 0 — — —
-38 lucida alba (4) Manettii (7) 13 0 — — —
-37 lucida alba (4) setipoda* (5) AAEE 23 0 — — —
49-38 lucida alba (4) suffulta (5) CCDD 15 20 8.7 47.2 18
42-38 lucida alba (4) xanthina allard (1) 15 13 3.5 47.2 7
40-39 lucida alba (4) xanthina allard (1) 9 22 9.0 47.2 19
68-38 multiflora* (6) AA foet. bicol. (1) BBDD 31 13 2.0 10.3 19
-38 mult. cathay. (6)AA foet. bicol. (1) BBDD 17 0 — — —
-39 mult. cathay. (6)AA foet. bicol. (1) BBDD 24 0 — — —
-39 mult. cathay. (6)AA gall. cond.* (2) AACC 28 0 — — —
66-38 mult. cathay. (6)AA Manettii (6) 20 5 2.0 10.5 19
-38 mult. cathay. (6)AA moschata flor. (6) 11 0 — — —
52-39 mult. cathay. (6)AA xanthina allard (1) 31 3 2.0 10.5 19
64-38 mult. chenault (6) foet. bicol. (1) BBDD 25 4 2.0 9.3 22
-36 mult. Iowa damas. trig.* (7) AACC 13 23 — — —
-36 mult. Iowa foot. bicol. (1) BBDD 9 11 — — —
-36 mult. Iowa gallica (2) AACC 8 0 — — —
-36 mult. Iowa lucida (4) CCDD 12 0 — — —
-36 mult. Iowa spinosissima* (1) BBCC 10 0 — — —
-39 mult. upright (6)AA foet. bicol. (1) BBDD 24 0 — — —
63-39 mult. upright (6)AA gall. cond.*(2) AACC 21 43 3.4 10.4 33
58-39 mult. upright (6)AA xanthina allard (1). 23 13 1.7 10.4 16

*Chromosome number of this form accepted as reported by others; not counted by us.
**No chromosome count made or reported on this form, but all counts for this species and its varieties suggest the indicated number.
†Number in parenthesis refers to generic section, after Rehder (1940), see table.
‡Letters refer to Hurst's septet classification.

TABLE I— (concluded)

Cross
No.
Seed Parent
(Diploid)
Pollen Parent
(Tetraploid)
No
Flowers
Crossed
Per
Cent
Set
Seeds Per Hip Per
Cent
Filled
Crossed O. P.
-36 mult. Welch (6)AA damas. trig.* (2) AACC 10 10 — — —
-36 mult. Welch (6)AA foetida* (1) BBDD 10 0 — — —
-36 mult. Welch (6)AA foet. bicol. (1) BBDD 8 0 — — —
-36 mult. Welch (6)AA gallica (2) AACC 18 0 — — —
-37 odorata (7) foet. bicol. (1) BBDD 53 0 — — —
-38 rugosa (5) CC feet. bicol. (1) BBDD 8 0 — — —
-39 rugosa (5) CC foet. bicol. (1) BBDD 12 0 — — —
-39 setigera A (6) suffulta* (5) 14 0 — — —
-39 setigera A (6) villosa recondita (3) 13 0 — — —
-39 setigera (6) AA xanthina allard (1) 15 0 — — —
-39 setigera (6) AA foet. bicol. (1) BBDD 15 0 — — —
-39 setigera (6) AA gall. cond. (2) AACC 15 0 — — —
-38 setigera (6) AA centifolia (2) AACC ca. 20 0 — — —
124-38 setigera (6) AA damas. trig.* (2) AACC ca. 20 15 5.7 4.9 116
119-38 setigera (6) AA foet. bicol. (1) BBDD ca. 20 5 1.0 4.9 20
-37 setigera (6) AA gallica (2) AACC 31 0 — — —
-38 setigera (6) AA gall. spl.** (2) AACC ca. 20 0 — — —
121-38 setigera (6) AA Manettii (7) ca. 20 10 10.0 4.9 204
-38 setigera (6) AA moschata flor. (6) ca. 20 0 — — —
-38 setigera (6) AA saturata* (5) CCDD ca. 20 0 — — —
73-37 setigera (6) AA setipoda* (5) AAEE 35 17 8.8 4.9 180
-38 setigera (6) AA Suffulta* (5) CCDD ca. 20 0 — — —
120-38 setigera (6) AA xanthina allard (1) ca. 20 10 2.0 4.9 41
-37 Soulieana (6) AA foet. bicol. (1) BBDD 30 0 — — —
83-37 Soulieana (6) AA foet. bicol. (1) BBDD 46 35 1.6 7.0 23
77-38 Soulieana (6) AA foet. bicol. (1) BBDD 18 32 1.2 7.0 17
-39 Soulieana (6) AA foet. bicol. (1) BBDD 20 0 — — —
77-37 Soulieana (6) AA gallica (2) AACC 41 59 5.1 7.0 73
78-37 Soulieana (6) AA gall. spl.** (2) AACC 30 37 3.8 7.0 54
-38 Soulieana (6) AA Manettii (7) 20 0 — — —
74-37 Soulieana (6) AA saturata* (5) CCDD 38 87 4.2 7.0 80
89-38 Soulieana (6) AA suffulta* (5) CCDD 19 53 2.2 7.0 31
85-38 Soulieana (6) AA xanthina allard (1) 24 38 4.0 7.0 43
-39 Souheana (6) AA xanthina allard (1) 20 0 — — —
-38 Wichuraiana (6) AA centifolia (2) AACC 21 0 — — —
106-38 Wichuraiana (8) AA damas. trig.* (2) AACC 17 35 2.7 23.85 11
97-38 Wichuraiana (6) AA foetida (1) BBDD 17 12 1.5 23.85 6
-39 Wichuraiana (8) AA foet. bicol. (1) BBDD 22 0 — — —
110-38 Wichuraiana (6) AA gall. spl.** (2) AACC 17 29 1.8 23.85 8
92-39 Wichucatana (6) AA gall. cond.* (7) AACC 20 5 1.0 23.85 4
99-38 Wichuraiana (8) AA Manettii (6) 19 18 2.7 23.85 11
-38 Wichuraiana (6) AA saturata (5) CCDD 17 0 — —  
111-38 Wichuraiana (6) AA suffulta* (5) CCDD 22 27 2.8 23.85 12
-39 Wichuraiana (6) AA suffulta* (5) CCDD 19 0 — — —
102-38 Wichuraiana (6) AA xanthina allard (1) 21 38 1.8 23.85 8
-39 Wichuraiana (6) AA xanthina allard (1) 20 0 — — —
-37 Willmottiae (5) BB foet. bicol. (1) BBDD 12 0 — — —
-15 Woodsii (1) DD foet. bicol. (5) BBDD 15 7 2.0 15.3 13

*Chromosome number of this form accepted as reported by others; not counted by us.
**No chromosome count made or reported on this form, but all counts for this species and its varieties suggest the indicated number.
†Number in parenthesis refers to generic section, after Rehder (1940), see table.
‡Letters refer to Hurst's septet classification.

TABLE II—CROSSES BETWEEN DIPLOID AND PENTAPLOID ROSE SPECIES

Cross
No.
Seed Parent
(Diploid)
Pollen Parent
(Pentaploid)
No.
Flowers
Crossed
Per
Cent
Set
Seeds Per Hip Per
Cent
Filled
Crossed O.P.
-39 Beggeriana (5)†DD‡ cor. Froe.** (3) ACDDE 12 0 — — —
-39 Beggeriana (5)†DD‡ morica 11 0 — — —
-39 Helenae (6) AA cor. Froe.** (3) ACDDE 13 0 — — —
44-38 lucida alba (4) canina (3) AABDE 14 14 21.5 47.2 5
-37 lucida alba (4) cor. Froe.** (3) ACDDE 31 0 — — —
-37 lucida alba (4) mundi 18 0 — — —
-38 lucida alba (4) mundi 13 0 — — —
-38 mult. cathay. (6)AA Alberti (5) 27 0 — — —
-38 mult. cathay. (6)AA canina* (3) AABDE 12 0 — — —
-39 mult. cathay. (6)AA canina* (3) AABDE 24 0 — — —
-39 mult. cathay. (6)AA cor. Froe.** (3) ACDDE 27 0 — — —
-39 mult. cathay. (6)AA ferox* (3) 37 0 — — —
-36 mult. Iowa (6) AA morica 7 0 — — —
-36 mult. Iowa (6) AA mundi 8 0 — — —
59-39 mult. upright* (6) AA canina* (3) AABDE 20 0 1.0 10.4 10
-36 mult. Welch (6)AA mundi 10 0 — — —
122-38 setigera (2) AA canina* (3) AABDE ca. 20 0 1.0 4.9 20
-38 setigera (2) AA morica ca. 20 0 — — —
-38 setigera (2) AA mundi ca. 20 0 — — —
-39 setigera C (6) AA cor. Froe.** (3) ACDDE 14 0 — — —
-39 setigera C (6) AA morica 13 0 — — —
-39 setigera A (6) mundi 13 0 — — —
-39 Soulieana (6) AA Alberti (5) 20 0 — — —
-39 Soulieana (6) AA banksiopsis (5) 20 0 — — —
86-38 Soulieana (6) AA canina* (3) AABDE 23 17 1.3 7.0 18
-39 Soulieana (6) AA aanina* (3) AABDE 21 0 — — —
-37 Soulieana (6) AA morica 37 0 — — —
-38 Soulieana (6) AA morica 19 0 — — —
-37 Soulieana (6) AA mundi 45 0 — — —
79-38 Soulieana (6) AA mundi 23 9 1.0 7.0 Il
85-39 Wichuraiana (6) AA Alberti (5) 17 65 4.0 23.83 17
104-38 Wichuraiana (6) AA canina* (3) AABDE 23 17 2.5 23.85 10
88-39 Wichuraiana (6) AA canina* (3) AABDE 17 35 4.5 23.85 19
90-39 Wichuraiana (6) AA cor. Froe.** (3) ACDDE 22 9 1.0 23.85 4
91-39 Wichuraiana (6) AA ferox* (3) 10 20 2.5 23.85 10
95-39 Wichuraiana (6) AA morica 22 5 3.0 23.85 13
113-38 Wichuraiana (6) AA mundi 17 24 1.5 23.85 6
-37 Willmottiae (5) BB morica 11 0 — — —
-37 Willmottiae (5) BB mundi 15 0 — — —

*Chromosome number of this form accepted as reported by others; not counted by us.
**No chromosome count made or reported on this form, but all counts for this species and its varieties suggest the indicated number. 
†Number in parenthesis refers to generic section, after Rehder (1940), see table.
‡Letters refer to Hurst's septet classification.

In roses the basic chromosome number is 7. In the crosses reported here, diploid, tetraploid, pentaploid, hexaploid and (probably) octoploid parents were used. Nineteen different combinations of chromosome numbers were attempted. Rosa cinnamomea was the only octoploid, or near-octoploid, species that was used, and exact counts on our plants of this form have not been made. Since the number is definitely very near 56, crosses involving this form are included for any possible additional information available from them.

Tables I to X inclusive record crosses attempted from 1936 through 1939. The crosses in these tables are arranged on the basis of respective chromosome numbers of seed and pollen parents. The left hand column of each table gives information relative to the number of the cross: if unsuccessful a number, as -36 or -39, records the year of the attempt; if successful a number has been assigned to the cross as 33-37 (see Table I, near top) which means this was the thirty-third successful cross to germinate in the year 1937. In the sixth column is found the average number of seeds per hip resulting from the con-trolled pollinations, while in the seventh Column appear the average number of seeds in from 10 to 20 open-pollinated fruits of the seed parent. Number of seeds per hip in open-pollinated fruits were secured in 1939 except for Rosa Beggeriana,  R. gallica, and R. Woodsii where 1940 counts were made. The figures in the last or "per cent filled" column are based on this average number of seeds in 1939 (or 1940) open-pollinated fruits as 100 per cent. This per cent is used as one measure of fertility for the various crosses. Doubtless the 100 per cent point of fill will show some, but very probably not marked, fluctuation from year to year.

TABLE III—CROSSES OF DIPLOID WITH HEXAPLOID AND OCTOPLOID ROSE SPECIES

Cross
No.
Seed Parent
(Diploid)
Pollen Parent No.
Flowers
Crossed
Per
Cent
Set
Seeds Per Hip Per
Cent
Filled
Crossed O. P.
2X x 6X
-39 Beggeriana (5)†DD‡ hibernica glabra (1) 14 0 — — —
-39 Beggeriana (5)†DD‡ nutkana (5) AADDEE 13 0 — — —
-39 mult. cathay. (6) AA hibernica glabra (1) 25 0 — — —
-36 mult. Iowa (6) AA nutkana (5) AADDEE 10 0 — — —
-36 mult. Welch (6) AA nutkana (5) AADDEE 16 0 — — —
-38 setigera (6) AA nutkana (5) AADDEE ca. 20 0 — — —
-39 setigera A (6) AA nutkana (5) AADDEE it 0 — — —
-39 setigera C (6) AA hibernica gravesii (1) 15 0 — — —
-39 setigera C (6) AA Macounii (5) 14 0 — — —
-37 Soulieana (6) AA nutkana (5) AADDEE 30 0 — — —
78-38 Soulieana (6) AA nutkana (5) AADDEE 18 11 1.0 7.0 14
93-39 Wichuraiana (6) AA hibernica glabra (1) 21 14 1.0 23.85 4
-39 Wichuraiana (6) AA hibernica gravesii (11) 19 0 — — —
94-39 Wichuraiana (6) AA Macounii (5) 19 11 1.5 23.85 6
100-39 Wichuraiana (6) AA nutkana (5) AADDEE 21 5 1.0 23.85 4
44-37 Woodsii (5) DD nutkana (5) AADDEE 36 47 18.5 15.3 121
2X x 8X (?)
-39 mult. cathay. (6)† cinnamomea (5) 25 0 — — —
-39 mult. cathay. (6) † cinnamomea (5) 28 0 — — —
87-38 Soulieana (6) cinnamomea (5) 17 12 1.5 7.0 21
-39 Soulieana (6) cinnamomea (5) 21 0 — — —
105-38 Wichuraiana (6) cinnamomea (5) 18 17 2.3 23.85 10
87-39 Wichuraiana (6) cinnamomea (5) 19 5 1.0 23.85 4

*Chromosome number of this form accepted as reported by others; not counted by us.
**No chromosome count made or reported on this form, but all counts for this species and its varieties suggest the indicated number.
†Number in parenthesis refers to generic section, after Rehder (1940), see table.
‡Letters refer to Hurst's septet classification.

As where diploids are crossed with diploids, some species set much more readily in interspecific crosses than others. Where diploid seed parents were crossed with polyploids in an appreciable number of cases Rosa Soulieana and R. Wichuraiana gave results above the average. For all diploids 46 per cent of species crosses and 14 per cent of pollinations set; for Soulieana these figures were 48 and 19.5 per cent respectively; and for Wichuraiana the per cents were 76 and 15. Among the tetraploid seed parents gallica (but not its varieties) was exceptional. For all tetraploid crosses per cents of crosses and of pollinations settings were 50 and 17 respectively; for gallica alone these figures were 79 and 37, much above the average in each case. The three pentaploid seed parents used all gave good results but ccinina was the best of the three in our crosses.

TABLE IV—CROSSES BETWEEN TETRAPLOID AND DIPLOID ROSE SPECIES

Cross
No.
Seed Parent
(Tetraploid)
Pollen Parent
(Diploid)
No
Flowers
Crossed
Per
Cent
Set
Seeds Per Hip   Per
Cent
Filled
Crossed O. P.
-38 bella* (5)† BBEE‡ mult. chenault* (6) AA 12 0 — — —
-38 bella* (5)† BBEE‡ mult. Welch (6) AA 12 0 — — —
-37 bella* (5)† BBEE‡ odorata (7) 17 0 — — —
-37 blanda A (5) Multiflora Iowa (6) 49 0 — — —
-37 centifolia (2) AACC odorata (7) 22 0 — — —
-37 damas. trig.* (7) AACC abyssinica (6) AA 25 0 — — —
-19 damas. trig.* (7) AACC gigantea* (7) AA 12 0 — — —
39-37 damas. trig.* (7) AACC lucida alba (4) 25 36 13.1 10.3 127
-37 damas. trig.* (7) AACC multiflora* (6) AA 35 0 — — —
-39 damas. trig.* (7) AACC mult. chenault (6) AA. 14 0 — — —
8-37 damas. trig.* (7) AACC mult. Iowa (6) AA 34 0 1.0 10.3 10
-39 damas. trig.* (7) AACC mult, upright (6) AA 12 0 — — —
-37 damas. trig.* (7) AACC odorata (7) 27 0 — — —
-39 damas. trig.* (7) AACC odorata (7) 13 0 — — —
66-39 damas. trig.* (7) AACC rugosa (5) CC 18 67 7.8 10.3 76
34-37 damas. trig.* (7)AACC Wichuraiana (6) AA 20 20 1.8 10.3 17
57-37 gallica (2) AACC Blanda (5) DD 21 81 28 25.0 104
24-38 gallica (2) AACC Fendleri* (5) DD 15 27 16.2 25.0 65
30-38 gallica (2) AACC Giraldii (5) EE 14 14 14.0 25.0 56
26-38 gallica (2) AACC Hugonis* (1) BB 16 38 13.8 25.0 55
?-38 gallica (2) AACC multiflora* (6) AA 19 58 29.7 25.0 119
-38 gallica (2) AACC mult. chenault (6) AA 14 0 — — —
61-37 gallica (2) AACC multiflora Iowa (6) AA 30 23 10.1 25.0 40
59-37 gallica (2) AACC multiflora Welch (6)AA 28 43 18.7 25.0 75
54-37 gallica (2) AACC odorata (7) 46 24 7.8 25.0 31
22-38 gallica (2) AACC pisocarpa (5) DD 16 31 33.4 25.0 134
55-37 gallica (2) AACC rugosa (5) CC 28 71 31.5 25.0 126
-38 gallica (2) AACC Willmottiae (5) BB 16 0 — — —
32-37 gallica (2) AACC Woodsii (5) DD 27 22 267 25.0 —
29-38 gallica (2) AACC xanthina (1) BB 13 46 13.8 25.0 55
-37 gall. spl.** (2) Giraldii (5) 23 0 — — —
-36 gall. spl.** (2) mult. cathay. (6) 9 0 — — —
-37 gall. spl.** (2) mult. Iowa (6) 27 0 — — —
-37 gall. spl.** (2) mult. platyphylla (6) 35 0 — — —
1-37 gall. spl.** (2) pisocarpa (5) 26 8 6.5 5.8 112
10-37 gall. spl.** (2) xanthina (1) 26 4 1 5.8 17
-38 lucida (4) CCDD abyssinica (6) AA 12 0 — — —
-38 lucida (4) CCDD mult. chenault (8) AA 14 0 — — —
2-37 lucida (4) CCDD mult. Iowa (6) AA 19 5 8. 37.4 21
-38 lucida (4) CCDD mult, upright (6) AA 16 0 — — —
-38 lucida (4) CCDD mult. Welch (6) AA 14 0 — — —
-38 lucida (4) CCDD odorata (7) 13 0 — — —
-37 lucida (4) CCDD rugosa (5) CC 29 0 — — —
-37 lucida (4) CCDD Wichuraiana (6) AA 22 0 — — —
-37 lucida CCDD xanthina BB 14 0 — — —
72-39 (4)  moschata flor. (6) (1)  mult. upright (6) 26 96 5.7 14.5 39
38-37 nitida (4) CC lucida alba (4) 19 26 3.2 33.8 9
-37 nitida (4) CC rugosa (5) 21 0 — — —
-37 oxyodon DDEE mult. chenault (6) AA 13 0 — — —
-37 oxyodon DDEE mult. Welch (6) AA 10 0 — — —
-37 oxyodon DDEE odorata (7) 14 0 — — —
-37 setipoda* (5) AAEE mult. cathay. (6) AA 49 0 — — —
-37 setipoda* (5) AAEE mult. Iowa (8) AA 28 0 — — —
-37 setipoda* (5) AAEE odorata (7) 32 0 — — —

*Chromosome number of this form accepted as reported by others; not counted by us.
**No chromosome count made or reported on this form, but all counts for this species and its varieties suggest the indicated number.
†Number in parenthesis refers to generic section, after Rehder (1940), see table.
‡Letters refer to Hurst's septet classification.

Table XI summarizes the crosses listed in Tables I to X, and in addition includes data from diploid by diploid crosses for completeness and comparison.

Effectiveness of Parents on Basis of Chromosome Grouping:— Comparing the figures of Table XI it would seem that on the basis of chromosome numbers the pentaploids are easily the best seed parents; while the tetraploids and cliploids are close together in value, with good setting percentages both of interspecific combinations and total pollinations. Where seed parents were pentaploid about 80 per cent of the crosses set, where they are tetraploid about 50 per cent set, and where diploid about 40 per cent set (with one exception), regardless of the number of chromosomes in the pollen. This consistency would seem to speak for the significance of the figures. Hexaploids and octoploids can not be included in this group comparison because of the small number of crosses. It may be pointed out, however, that where hexaploid Rosa nutkana (Table IX) and octoploid R. cinnamomea (Table X) were used as seed parents in attempted crosses, setting percentages were very low.

TABLE V—CROSSES BETWEEN TETRAPLOID ROSE SPECIES

Cross
No.
Seed Parent
(Tetraploid)
Pollen Parent
(Diploid)
No
Flowers
Crossed
Per
Cent
Set
Seeds Per Hip Per
Cent
Filled
Crossed O.P.
-39 blanda A (5)† foet. bicol. (1) 23 0 — — —
48-37 blanda A (5)† oxyodon 41 49 8.8 8.06 107
l-39 blanda A (5)† xanthina allard (1) 12 0 1.0 8.06 12
-38 bella (5) BBEE‡ damas. trig.* (2)AACC 11 0 — — —
-37 bella (5) BBEE‡ foet. bicol. (1) BBDD 17 0 — — —
-38 bella (5) BBEE‡ Manettii (6) 13 0 — — —
-37 centifolia (2) AACC foet. bicol. (1) BBDD 19 0 — — —
11-37 damas. trig.* (2) AACC foet. bicol. (1) BBDD 17 12 1.0 10.3 10
-39 damas. trig.* (2) AACC foet. bicol. (1) BBDD 13 0 — — —
-39 damas. trig.* (2) AACC xanthina allard (1) 13 0 — — —
56-37 gallica (2) AACC foet. bicol. (1) BBDD 32 22 10.4 25.0 40
-38 gallica (2) AACC foet. bicol. (1) BBDD 17 0 — — —
60-37 gallica (2) AACC oxyodon DDEE 21 67 23.0 25.0 92
62-37 gallica (2) AACC saturata*(5) CCDD 29 69 33.4 25.0 134
27-38 gallica (2) AACC suffulta (5) CCDD 19 42 28.2 25.0 113
28-38 gallica (2) AACC xanthina allard (1) 13 54 7.3 25.0 —
51-38 lucida (4) CCDD foet. bicol. (1) BBDD 12 8 2.0 37.4 5
-37 lucida (4) CCDD gallica (2) AACC 28 0 — — —
-38 lucida (4) CCDD Manettii (7) 15 0 — — —
37-37 lucida (4) CCDD setipoda* (5) AAEE 26 85 34.5 37.4 92
-38 lucida (4) CCDD xanthina allard (1) 15 0 — — —
-39 moschata flor. (6) foet. bicol. (1) BBDD 29 0 — — —
-39 moschata flor. (6) moschata flor. (6) 53 0 — — —
74-39 moschata flor. (6) xanthina allard (1) 27 7 1.0 14.5 7
-37 nitida (4) gallica (2) 13 0 — —  
31-37 nitida (4 setipoda* (5) 23 35 14.9 33.8 40

*Chromosome number of this form accepted as reported by others; not counted by us.
†Number in parenthesis refers to generic section after Rehder (1940), see table.
‡Letters refer to Hurst's septet classification.

TABLE VI—CROSSES OF TETRAPLOID WITH PENTAPLOID, HEXAPLOID AND OCTOPLOID (?) ROSE SPECIES

Cross
No.
Seed Parent
(Tetraploid)
Pollen Parent No.
Flowers
Crossed
Per
Cent
Set
Seeds Per Hip Per
Cent
Filled
Crossed O. P.
4X x 5X
-39 damas. trig.*(2)†AACC canina* (3) AABDE 14 0 — — —
-39 damas. trig.*(2)†AACC cor. Froe.**(3)ACCDE 7 0 — — —
-38 gallica (2) morica 14 7 2.0 25.0 8
-38 gallica (2) mundi 17 30 8.0 25.0 32
-38 lucida (4) CCDD canina* (3) AABDE 13 0 — — —
5-37 Lucida (4) CCDD coriifolia* (3) ACDDE 20 5 4.0 37.4 11
70-39 damas. trig. (2) AACC nutkana (5) AADDEE 15 80 10.5 10.3 102
4X x 6X
58-37 Gallica (2) AACC nutkana (5) AADDEE 31 58 15.7 25. 63
4X x 8X (?)
-38 gallica (2) Cinnamomea (5) 16 0 — — —

*Chromosome number of this form accepted as reported by others; not counted by us.
**No chromosome count made or reported on this form, but all counts for this species and its varieties suggest the indicated number.
†Number in parenthesis refers to generic section, after Rehder (1940), see table.
‡Letters refer to Hurst's septet classification.

TABLE VII—CROSSES BETWEEN PENTAPLOID AND DIPLOID ROSE SPECIES

Cross
No.
Seed Parent
(Pentaploid)
Pollen Parent
(Diploid)
No
Flowers
Crossed
Per

Cent Set

Seeds Per Hip Per
Cent
Filled
Crossed O.P.
16-39 canina (3)† AABDE blanda** (5) 14 43 2.2 21.8 10
17-39 canina (3) AABDE‡ fendleri (5) DD 16 13 1.0 21.8 5
19-39 canina (3) AABDE‡ gigantea (7) AA 12 83 8.9 21.8 41
-39 canina (3) AABDE‡ Giraldii (5) EE 12 0 — — —
73-38 canina (3) AABDE‡ mult. chenault (6)  AA 12 8 3.0 21.8 14
21-39 canina (3) AABDE‡ mult. chenault (6) AA 14 29 1.8 21.8 8
-38 canina (3) AABDE‡ mult. upright (6) AA 10 0 — — —
6-39 canina (3) AABDE‡ mult, upright (6) AA 16 44 3.7 21.8 17
7-39 canina (3) AABDE‡ odorata (7) 22 77 2.5 21.8 11
-38 canina (3) AABDE‡ odorata (7) 11 0 — — —
13-39 canina (3) AABDE‡ rugosa (5) CC 26 73 10.4 21.8 48
-38 cor. froe.** (3) ACDDE hugonis* BB (1) 11 — — — —
36-37 cor. froe.** (3) ACDDE lucida alba (4) 21 67 12.6 20.3 62
27-37 cor. froe.** (3) ACDDE mult. cathay. (6) AA 13 15 4.0 20.3 20
-38 cor. froe.** (3) ACDDE mult. chenault (6) AA 12 0 — — —
24-37 cor. froe.** (3) ACDDE mult. Iowa (6) AA 17 6 5.0 20.3 25
34-38 cor. froe.** (3) ACDDE mult. Iowa (6) AA 11 18 6.5 20.3 32
-38 cor. froe.** (3) ACDDE mult. Welch (6) AA 12 0 — — —
3-37 cor. froe.** (3) ACDDE odorata (7) 28 20 3.0 20.3 15
31-38 cor. froe.** (3) ACDDE odorata (7) 11 9 1.0 20.3 5
26-37 cor. froe.** (3) ACDDE rugosa (5) CC 9 22 3.0 20.3 15
33-38 cor. froe.** (3) ACDDE rugosa alba (5) CC 12 25 1.3 20.3 6
14-37 cor. froe.** (3) ACDDE Wichuraiana (6) AA 19 16 13.7 20.3 67
-38 cor. froe.** (3) ACDDE xanthina (1) BB 13 0 — — —
-37 Rubiginosa (3) ABBCD Abyssinica (6) AA 19 0 — — —
34-39 rubiginosa (3) ABBCD gigantea (7) AA 17 29 14.4 27.5 52
21-37 rubiginosa (3) ABBCD mult. cathay. (6) AA 19 21 23.2 27.5 84
26-39 rubiginosa (3) ABBCD mult. cathay. (6) AA 21 95 11.2 27.5 41
27-39 rubiginosa (3) ABBCD mult. chenault (6) AA 33 13 2.7 27.5 10
-37 rubiginosa (3) ABBCD mult. Iowa (6) AA 24 0 — — —
28-89 rubiginosa (3) ABBCD mult, upright (6) AA 17 47 4.4 27.5 16
-37 rubiginosa (3) ABBCD odorata (7) 24 0 — — —
30-39 rubiginosa (3) ABBCD odorata (7) 20 50 3.4 27.5 12
22-37 rubiginosa (3) ABBCD Wichuraiana (6) AA 31 13 24 27.5 87

*Chromosome number of this form accepted as reported by others not counted by us.
**No chromosome count made or reported on this form, but all counts for this species and its varieties suggest the indicated number.
† Number in parenthesis refers to generic section, after Rehder (1940) see table.
‡Letters refer to Hurst's septet classification.

If a significant per cent of maternals should appear among the future progeny from pentaploid seed parents, as might be expected from the work of Tackholm and of Hurst these seed parent ratings would need revision. On the basis of present data, however, pentaploids must be considered the outstanding seed parent chromosome group.

Thompson (11) and others have pointed out the interesting observation that in most plant genera interspecific crosses between parents with different chromosome numbers are more apt to be successful if the species with the higher chromosome number is used as the seed parent than in the case of the reciprocal crosses. The nature of the main aim of the present problem has resulted in only a small number of actual reciprocal species crosses in the present instance. However it is desirable from the practical plant breeding standpoint of securing adequate numbers of seedlings from which to select desirable material for testing, to know if the relative size of the chromosome number of the two parents has any general tendency to influence results along this line.

By using the per cents in Table XI we may compare chromosome group crosses where the seed parents have identical, larger or smaller chromosome numbers than the pollen parents. If we omit the results of hexaploid by diploid crosses, which involved only five combinations, it is seen that in general where the seed parents had larger numbers of chromosomes there was a greater percentage of both crosses and pollinations setting, than in the other cases. But since it is apparent that this is clue here to the "weight" of the number and fertility of crosses with pentaploid seed parents, we must conclude that it is not a significant fact, and that there are apparent no important effects clue to proportional chromosome numbers of the parents.

TABLE VIII—CROSSES OF PENTAPLOID WITH TETRAPLOID, PENTAPLOID,
HEXAPLOID AND OCTOPLOID (?) ROSE SPECIES

Cross
No.
Seed Parent
(Pentaploid)
Pollen Parent
(Diploid)
No
Flowers
Crossed
Per
Cent
Set
Seeds Per Hip Per
Cent
Filled
Crossed O.P.
5X x 4X
75-38 canina (3)† AABDE‡ foet. bicol. (1) BBDD 11 18 2.0 21.8 9
5-39 canina (3)† AABDE‡ foet. bicol. (1) BBDD 10 32 1.7 21.8 8
18-39 canina (3)† AABDE‡ gall. cond.* (2) AACC 12 100 10.5 21.8 48
74-38 canina (3)† AABDE‡ Manettii (7) 13 89 4.0 21.8 18
9-39 canina (3)† AADDE‡ xanthina allard (1) 15 20 1.7 21.8 8
12-37 cor. froe.** (3) ACDDE foet. bicol. (1) BBDD 14 21 1.3 20.3 6
32-38 cor. froe.** (3) ACDDE foet. bicol. (1) BBDD 12 8 1.0 20.3 3
-37 cor. froe.** (3) ACDDE gallica splendens** (2) 17 0 — — —
37-38 cor. froe.** (3) ACDDE suffulta (5) CCDD 11 55 12.8 20.3 63
36-38 cor. froe.** (3) ACDDE xanthina allard (1) 13 15 3.5 20.3 17
23-37 rubiginosa (3) ABBCD foet. bicol. (1) BBDD 25 12 6.3 27.5 23
31-39 rubiginosa (3) ABBCD foet. bicol. (1) BBDD 21 43 3.2 27.5 12
-39 rubiginosa (3) ABBCD xanthina allard (1) 21 0 — — —
5X x 5X
15-39 canina* (3) AABDE Alberti (5) 16 50 12.1 21.9 96
20-39 canina* (3) AABDE morica 13 69 — — —
23-39 canina* (3) AABDE mundi 14 71 8.1 21.8 37
32-39 rubiginosa* (3) ABBCD canina* (3) AABDE 23 70 18.6 27.5 68
-39 rubiginosa* (3) ABBCD cor. froe.** (3) ACDDE 18 0      
5X x 6X
22-39 canina (3) AABDE hibernica glabra (1) 13 54 2.0 21.8 9
25-39 canina (3) AABDE nutkana (5) AADDEE 12 92 18.0 21.8 83
5X x 8X
35-38 cor. froe.** (3) cinnamomea (5) 12 50 4.3 20.3 21
33-39 rubiginosa (3) cinnamomea (5) 15 87 23.3 27.5 85

*Chromosome number of this form accepted as reported by others; not counted by us.
**No chromosome count made or reported on this form, but all counts for this species and its varieties suggest the indicated number.
†Number in parenthesis refers to generic section, after Rehder (1940), see table.
‡Letters refer to Hurst's septet classification.

There is apparently less difference in the effectiveness of forms in the various groups as pollen parents, than as seed parents, although it is probably more difficult to evaluate species on a basis of pollen potency. Indications from Table XI are that tetraploids are most potent as a pollen group, while pentaploids, especially in per cent of pollinations setting, are least effective. Both of these indications are borne out, and somewhat strengthened, by considering the results of all crosses; that is, those including all combinations of these pollen groups with hexaploid and octoploid seed parents, as well as those represented by the data of Table XI. The small number of crosses, as where seed parent groups were considered, does not allow a comparative rating for pollen potency of the two high polyploid groups. Reference to Tables III, VI, VIII and IX, however, will show that per cents of species combinations and of pollinations setting with hexaploid and octoploid pollen were high, and suggest that such forms may be especially valuable as pollen parents; additional crosses involving these chromosome forms are needed to confirm this.

TABLE IX—CROSSES OF HEXAPLOID WITH DIPLOID, TETRAPLOID, AND OCTOPLOIO (?) ROSE SPECIES

Cross
No.
Seed Parent
(Pentaploid)
Pollen Parent
(Diploid)
No
Flowers
Crossed
Per
Cent
Set
Seeds Per Hip Per
Cent
Filled
Crossed O. P.
6X x 2X
-38 nutkana(5)†AADDEE‡ Hugonis (1)BB 18 0 — — —
-38 nutkana(5)†AADDEE‡ multiflora(8) AA 17 0 — — —
-38 nutkana(5)†AADDEE‡ mult. chenault (6) AA 13 0 — — —
-37 nutkana(5)†AADDEE‡ mult. Iowa(6) AA 14 0 — — —
-38 nutkana(5)†AADDEE‡ mult. Iowa(6) AA 18 0 — — —
-37 nutkana(5)†AADDEE‡ mult. plat.(6) AA 15 0 — — —
-38 nutkana(5)†AADDEE‡ odorata (7) 21 0 — — —
-38 nutkana(5)†AADDEE‡ rugosa (5) CC 17 0 — — —
-37 nutkana(5)†AADDEE‡ Woodsii (5)DD 19 0 — — —
6X x 4X
-38 nutkana (5) foet. bicol. (1)BBDD 13 0      
-38 nutkana (5) saturata (5) CCDD 10 0      
-38 nutkana (5) xanthina allard (1) 13 0      
6X x 8X (?)
-38 nutkana (5) Cinnamomea (5) 15 7 21.0 8.8 238

†Number in parenthesis refers to generic section, after Rehder (1940), see table.
‡Letters refer to Hurst’s septet classification.

 TABLE X—CROSSES OF OCTOPLOID (?) WITH DIPLOID AND TETRAPLOID ROSE SPECIES

8X (?) x 2X
-38 cinnamomea (5)† mult. chenault (6) 13 0 — — —
-38 cinnamomea (5)† mult. Iowa (6) 19 0 — — —
-38 cinnamomea (5)† odorata (7) 17 0 — — —
-38 cinnamomea (5)† rugosa (5) 14 0 — — —
8X (?) x 4X
-38 cinnamomea (5) foetida bicolor (1) 17 0 — — —

†Number in parenthesis refers to generic section, after Rehder (1940), see table.

TABLE XI—SUMMARY OF CROSSES WITH PARENTS GROUPED ACCORDING TO CHROMOSOME NUMBERS,
SHOWING PER CENTS OF SPECIES COMBINATIONS AND OF POLLINATIONS SETTING*

  2x 4x 5x 6x
  Comb. Poll. Comb. Poll. Comb. Poll. Comb. Poll.
2x 43 16 61 13 38 5 42 8
4x 49 14 50 20 50 8 — —
5x 78 27 80 27 80 51 — —
6x 0 0 — — — — — —

*Only cases in which there were five or more species combinations made are considered here.

Comparison of Results of Diploid Crosses With Crosses Involving Polyploids:—In crosses between diploids alone (7) 94 species combinations and 4,755 pollinations were attempted, with about a fourth more of each, or 118 and 6,008 respectively, where polyploids were used. Species combinations setting for all crosses was 48.5 per cent, for diploid crosses alone was 43 per cent, and for all crosses with polyploicls was 54 per cent. Where pentaploicis were eliminated these per cents were 45.5, 43, and 48 respectively. In per cent of pollinations setting, the diploid crosses were slightly above, and the polyploid crosses somewhat below the set for all crosses. For of all crosses 15.5 per cent, of diploid crosses 16 per cent, and of all polyploid crosses 15 per cent of pollinations gave sets. Where pentaploids were eliminated these per cents were 14, 15, and 12 respectively. The deviations of per cents for the two different groups of crosses from that for all crosses appear to be small enough to he negligible. The same may he said for differences in setting per cents where pentaploids are included and excluded from the data. Tests of statistical significance would appear to be meaningless here because of the number and nature of the factors affecting the results.

2. Results Considered on Basis of Taxonomic Relations of Crossed Species

The arrangement of species by generic sections set forth in Rehder's "Manual of Cultivated Trees and Shrubs" (8) has been followed here, as in our previous paper. In Tables I to X the number in parenthesis following a species name refers to the taxonomic section to which Rehder (8) refers the species in question. The names of the generic sections corresponding to these numbers may be found in Table XII.

TABLE XII—REHDER'S SECTIONS OF ROSA

Second
Edition
(1940)
First
Edition
(1927)
Section Second Edition
(1940)
— 1 Simplicifoliae* Subgenus Hulthemia
1 11 Pimpinellifoliae Subgenus Eurosa
1 12 Luteae* Combined with Pimpinellifoliae
1 13 Sericeae* Combined with Pimpinellifoliae
2 7 Gallicanae Subgenus Eurosa
3 8 Caninae Subgenus Eurosa
4 9 Carolinae Subgenus Eurosa
5 10 Cinnamomeae Subgenus Eurosa
6 2 Synstylae Subgenus Eurosa
7 6 Indicae Subgenus Eurosa
8 3 Banksianae Subgenus Eurosa
9 5 Laevigatae Subgenus Eurosa
10 4 Bracteatae Subgenus Eurosa
— 14 Minutifoliae* SubgenusHesperhodos
— 15 Microphyliae* SubgenusPlatyrhodon

*These sectional names not used in 1940 edition.

For completeness this table also lists the numbers of the sections as given in the first edition, and the subgeneric names and inclusions as presented in the last edition, of Rehder's work, Rosa morica, R. mundi, and R. oxyodon,  which were used by us, are not included in Rehder's list. Data from these species is omitted from consideration in this section.

In Table XIII the partially summarized crossing data are arranged by numbers of the generic sections. The per cents of total species crosses and of total pollinations giving sets, and the per cent of fill of crossed hips — compared with open-pollinated hips — may all three be used as measures of crossing success in given individual crosses or groups of crosses,

From Table XIII it is apparent that crosses have been made between species within four different sections, i.e., Caninae, Carolinae, Cinnamomeae and Synstylae. A glance at the columns indicating fertility by per cents of set and fill suggest that crosses between species in the same section are apt to be more successful than crosses between species in different sections. This is in agreement with our conclusion based on crosses among diploids. Thus where two crosses were made between species in the Caninae one of the two set, a good per cent of the pollinations set, and a better fill of set hips was realized than in any other crosses where Caninae species were used as seed parents.

TABLE XIII—CROSSES PARTIALLY SUMMARIZED WITH SEED PARENT GROUPS ARRANGED ACCORDING TO REHDER'S SECTIONS

Sections
Crossed
No. of Sp.
Crosses
No.
Giving
Sets
Per
Cent
Setting
No. of
Pollinations
No. of
Sets
Per
Cent
Setting
Avg. Seeds Per Hip Per
Cent
Filled
Crossed O.P.
Gallicanae x
  Pimpinellifoliae 7 5 71 179 29 16 10.0 20.5 49
Caninae 2 0 0 21 0 0 — — —
Carolinae 1 1 100 25 9 36 13.1 10.3 127
Cinnamomeae 13 11 85 311 124 40 23.1 20.0 111
Synstylae 4 3 75 302 35 12 18.0 19.1 94
Indicae 4 1 25 120 11 9 7.8 25.0 31
Caninae x
  Pimpinellifoliae 9 6 67 187 40 21 6.0 22.4 31
Gallicanae 2 1 50 29 12 41 10.5 21.8 48
Caninae 2 1 50 41 16 39 18.6 27.5 68
Carolinae 1 1 100 21 14 67 12.6 20.3 62
Cinnamomeae 9 8 89 142 66 46 11.0 21.7 M
Synstylae 6 5 83 290 59 20 9.5 23.6 40
Indicae 6 6 100 155 57 37 8.1 22.5 23
Carolinae x
  Pimpinellifoliae 5 3 60 92 8 9 5.0 44.7 11
Gallicanae 4 2 50 90 2 2 2.0 47.2 4
Caninae 4 2 50 78 3 4 15.7 42.3 37
Carolinae 1 1 100 19 5 26 3.2 33.8 9
Cinnamomeae . 7 4 57 157 41 26 25.4 41.4 61
Synstylae 3 3 33 97 t 1 8.0 37.4 21
Indicae 3 0 0 41 0 0 — — —
Cinnamomeae x
  Pimpinellifoliae 15 3 20 265 5 2 5.2 16.5 32
Gallicanae 3 1 33 67 18 27 22.4 22.0 102
Caninae 1 0 0 12 0 0   — —
Cinnamomeae . 9 4 44 241 30 12 22.2 14.2 156
Synstylae 5 0 0 259 0 0 — — —
Indicae 5 0 0 100 0 0 — — —
Synstylae x
  Pimpinellifoliae 16 12 75 682 66 10 2.0 12.8 16
Gallicanae 10 6 60 350 83 18 4.0 14.4 28
Caninae 11 6 55 296 20 7 2,6 16.8 16
Carolinae 1 0 0 12 0 0 — — —
Cinnamomeae 20 10 50 575 77 13 3.8 17.7 21
Synstylae 4 1 25 110 25 23 5.7 14.5 39
Indicae 5 4 80 89 7 8 4.4 12.4 35
Indicae x
  Pimpinellifoliae. 1 0 0 53 0 0 — — —

In the single cross in which both species were in the Carolinae (Rosa nitida x R. lucida alba, Table IV) 26 per cent of the pollinations gave sets. This per cent was equalled in only one case in this seed group — where species of the Carolinae were pollinated by species of the Cinnamomeae — and it was much better than in other cases where the pollen parents were from different sections. The per cent of fill of the hips in this case was low, however.

Nine crosses were made between species belonging to the Cinnamomeae, and sets were secured from four of these. In only one case was the percentage of crosses giving sets exceeded when pollen parents from without the section were used, and here the excess was but slight. Per cent of pollinations giving sets, within Cinnamomeae, was low, but the per cent of fill of hips set was very high, much exceeding that of open pollinated hips.

Where crosses were made between species within the Synstylae it is seen that a larger per cent of pollinations have set, and a higher per cent of fill secured, than where crosses are made between seed species of this and pollen species of other sections. This is true even though only one cross of four attempted resulted in sets. The fact that the cross Rosa moschata floribunda x R. multiflora upright was so successful (Table IV) makes the high per cents in the last "per cent set" and the "per cent filled" columns appear to have a greater importance than is probably the actual case.

Value of Different Sections as Seed Parent Groups:— Table XIV is a summary of crosses with the seed parents in the different taxonomic sections being grouped. There are several points that have to be considered in evaluating the sections as seed parental groups. For one thing, as indicated by footnote three, since only one cross was involved, with the Indicae its inclusion here should not necessarily suggest it to be the least desirable seed parent section. Also, the large per cent of fill where species of the Cinnamomeae were seed parents should be somewhat discounted because a larger number of pollen parents of this same section were used than in the case of other sections.

TABLE XIV—SUMMARY OF DATA WITH SEED PARENTS GROUPED IN TAXONOMIC SECTIONS

Sections of
Seed Parent
No. of
Sections
of Pollen
Parent
No. of
Sp.
Crosses
No.
Giving
Sets
Per
Cent of
Crosses
Setting
No.
of
Pollinations
No.
of
Set
Per
cent
Setting
Per
cent
Filled
Gallicanae 7 (+3)* 34 24 71 1010 228 23 86
Caninae 8 (+2)* 37 30 81 892 283** 32 41
Carolinae 8 (+l)* 28 13 46 605 60 10 46
Cinnamomeae 7 (+3)* 43 9 21 1045 73 7 123
Synstylae 8 (+2)* 75 42 56 2366 265 11 22
Indicae 1 0 0 53 0 0 —

*Figure in parenthesis refers to crosses with number of species not included by Rehder and of unestablished subgeneric section.
**274 sets used here in computing per cent filled, due to lack of seed number data in 9 cases.
†lnsufficient for significance, included for completeness.

When these points are considered together with the data an approximate evaluation as seed parental groups of the various sections used in our crosses would place Gallicanae and Caninae at the top as about equally desirable on the face of things, but with the crossing value of the latter possibly having to be discounted because of chance of apomictic effect in some cases; somewhat less valuable in our crosses were the three sections Carolinae, Cinnamomeae and Synstylae, all of about equal value, everything considered. Indicae can not be rated here on the basis of the present meager data with it.

3. Results Considered on Basis of Hurst's Grouping of Species

TABLE XV—SUMMARY OF DATA WITH POLLEN PARENTS GROUPED IN TAXONOMIC SECTIONS

Sections of
 Seed Parent
No. of
Sections
of Pollen
Parent
No. of
Sp.
Crosses
No.
Giving
Sets
Per
Cent of
Crosses
Setting
No.
of
Pollinations
No. of
Sets
Per cent
Setting
Per
cent
Filled
Pimpinellifoliae 6 53 29 55 1458 148 10 28
Gallicanae 4 19 10 53 536 95 28 31
Caninae 5 20 9 45 448 39 9 35
Carolinae 4 4 3 75 77 28 36 52
Cinnamomeae 5 58 87 64 1426 338 24 71
Synstylae 5 (+1)* 23 10 43 1081 120 11 47
Indicae 5 (+1) 24 11 46 519 75 14 27

*Figure in parenthesis refers to crosses with number of species not included by Rehder and of unestablished subgeneric section.

Hurst (4, 5) considers similar species to be geographical subspecies of a single empirical, or "differential septet", species. Subspecies of one of these species, according to the theory, all exhibit similar botanical characters due to the presence of the same septet, or septet combination, of chromosomes. There are five such basic sets, with a resultant five, and only five, diploid differential septet species. Hurst believes that evolution in roses has been by descent from an original decaploid ancestor and that the five diploid "genetic" species represent the ultimate in the rose genus, with a possible 206 additional, polyploid septet combinations, giving that many possible polyploid species of roses existing now or having existed in the past. His studies enabled him to assign definite septet formula to a number of the species used here. These have been indicated (by capital letters) in Tables I to X. Presumably the species most closely related genetically will cross with the greatest ease, barring complications such as the presence of incompatibility factors.

Consequently it has seemed advisable to arrange our data so that general crossing phenomena may be observed among the groups with the different septet formulae. This arrangement has been made in Table XVI, and suggests some interesting genetic relationships.

TABLE XVI—COMPATIBILITY OF SPECIES AS ARRANGED ACCORDING TO HURST'S SYSTEM

  No. of
Comb.
Per
Cent
Setting
No. of
Poll.
Per
Cent
Setting
Per
Cent
Fill
AA x AA 28 46 756 14 36
  BB 15 60 375 6 13
  CC 7 57 220 5 12
  DD 17 53 346 10 28
  BE 4 50 121 2 14
  AACC 17 53 350 18 32
  AAEB 1 100 35 17 180
  BBCC 1 0 10 0
  BBDD 13 54 374 8 15
  CCDD 7 43 167 29 29
  AABDE 6 83 139 12 20
  ACDDE . 4 25 76 9 4
  AADDEE 6 33 126 2 6
AACC x AA 11 45 243 14 94
  BB 3 67 45 27 55
  CC 2 100 43 70 133
  DD 3 100 52 50 104
  EE 1 100 14 14 56
  BBDD 3 67 85 11 47
  CCDD 2 100 48 58 128
  DDEE 1 100 21 67 92
  AABDE 1 0 14 0
  ACCDE 1 0 7 0
  AADDEE 2 100 46 65 77
AAEE x AA 2 0 77 0
AABDE x AA 3 100 64 34 26
  CC 1 100 28 73 48
  DD 1 100 16 13 5
  EE 1 0 12 0
  AACC 1 100 12 100 48
  BBDD 1 100 30 27 8
  AADDEE 1 100 12 92 83
ABBCD x AA 7 71 171 26 44
  BBDD 2 100 46 26 15
  AABDE 1 100 23 70 68
  ACDDE 1 0 18 0
ACDDE x AA 6 67 84 10 41
  BB 2 0 24 0
  CC 2 100 21 24 10
  BBDD 4 100 52 15 6
  CCDD 1 100 11 55 63
 
  No. of
Comb.
Per
Cent
Setting
No. of
Poll.
Per
Cent
Setting
Per
Cent
Fill
AADDEE x AA 4 0 77 0
  BB 1 0 18 0
  CC 1 0 17 0
  DD 1 0 19 0
  BBDD 1 0 13 0
  CCDD 0 0 10 0
BB x AA 1 100 15 7
  BBDD 1 0 12 0
BBEE x AA 2 0 24 0
  AACC 1 0 Il o
  BBDD 1 0 17 0
CC x AA 1 0 13 0
  CC 1 0 21 0
  DD 1 100 11 9 69
  BBDD t 0 12 0
CCDD x AA 6 17 97 1 21
  BB 1 0 14 0
  CC 1 0 29 0
  AACC 1 0 28 0
  AAEE 1 100 26 85 92
  BBDD 1 100 12 8 5
  AABDE 1 0 13 0
  ACDDE 1 100 20 5 11
DD x AA 3 100 88 31 51
  BB 5 100 94 43 80
  CC 2 100 38 79 114
  DD 6 100 135 68 87
  EE 1 100 36 56 93
  AACC . 1 100 29 62 102
  BBDD 3 67 65 6 26
  CCDD 1 100 45 24 125
  CDDE 1 0 12 0
  AADDEE 2 50 49 35 121
DDEE x AA 2 0 23 0
EE x CC 1 0 26 0
  DD 1 0 47 0
  AACC 1 0 27 0
  BBDD 1 0 26 0
  CCDD 1 100 72 1
  DDEE 1 0 23 0

Where two or more septet combinations have been made, all seed groups with one or two A septets, except the hexaploid AADDEE, have given good percentages of set and fill. The third, fifth and sixth columns, Table XVI, bring this out. It may be observed however that where A and C septets are combined as in the seed groups AACC, ABBCD, and ACDDE that percentage of both set and fill is increased in the majority of cases where the same pollen group is concerned, over that of seed groups possessing A septets without C, or vice versa. The seed group AABDE, with comparatively few species combinations, is somewhat of an exception in that the set here was as good as in AC combination. In addition, where two A septets are present in both the seed and the pollen parents the accelerating effect of C, when present in the former, is not so obvious. This increased compatibility may be due to acceleration of pollen tube growth or some related phenomena. It is also apparent from the table that where seed groups with the formula DD have been used that compatibility and fertility, with various pollen groups, is good.

Numbers of combinations and pollinations with the remaining seed groups are probably not large enough to give much significance to other observed trends, such as the apparent depressing effect of C on DD septets, so far as compatibilities and fertilities of various groups are concerned. Some positive tendencies and affinities are also indicated with certain of the septet combinations in the pollen groups, but these are not brought out definitely enough by present data to warrant discussion at this time.

The data presented in Table XVI have been checked to see if crosses involving polyploids in which the pollen contained septets identical with those of the seed parent were any more prolific than where such septets were different in the different parents. When the data are considered either separately (as in Table XVI), or are combined and averaged, fertility is apparently not increased by pollinating with forms having (one or more) septets corresponding to those reported for the seed parent (although results from use of pollen carrying AA septets on seed forms with this group suggests an exception).

A partial explanation of the productive and fertile results of seed species with septet formulae AA, AACC, and DD may be furnished by the data on which Table XVII is based. This table, which includes only the species which Rehder (1940) describes in detail, is substantiated by available data from all species. It shows that the Gallicanae is mainly a tetraploid, AACC section; and that the Synstylae is essentially a diploid, AA section. Also, while not shown so clearly by the table, the Carolinae is predominantly a diploid DD section; four of the five species have chromosome numbers of 14 (28 has also been reported for two of these same ones), and the four with given septet formula have reported DD septets alone or in combination. The Caninae, composed of the cytologically irregular species, is made tip principally of pentaploid forms with some tetraploid and hexaploid ones, having all of Hurst's septets represented among the species but with A septets occurring most frequently especially in duplicate condition. Members of the Caninae used as seed parents in our crosses possessed either AA or AC septets, in combination with others. In the other sections there is a more even distribution of chromosome numbers, of septet occurrence, or of both — as is decidedly the case in the Cinnamomeae.

TABLE XVII—OCCURRENCE OF CHROMOSOME NUMBERS AND OF HURST'S SEPTETS
IN DIFFERENT TAXONOMIC SECTIONS

Taxonomic
Section
No. of Species
Described by
Rehder 1940
No. of These Sp.
on Which Data
are Available
Chromosome
Numbers†
Septets
Represented†
Pimpinellifoliae 7 5 28,14 BB*, DD, CC
Gallicanae 4 4 28‡ AACC*
Caninae 8 8 35**, 18 A, D, E, B, C
Carolinae 5 5 14**, 28‡ DD*, AA, CC
Cinnamomeae 25 24 14, 28, 42, 56 DD, EE, BB, AA, CC
Synstylae 13 12 14‡ AA*
Indicae 4 2 14-42 AA

†Where more than one number or one septet combination has been reported for species of a section, such numbers or septets are arranged, from left to right, in order of their frequency of occurrence.
‡One form in which this number occurs has also been reported as having a higher number.
*A preponderance of the species possess these septets.
**A preponderance of the species have this chromosome number.

It appears significant that the septet formulae possessed by the majority of the most prolific and fertile seed parent species (Rosa gallica, R. Soulieana, R. wichuraiana, and probably including the pentaploid Caninae species, but with fertile R. blanda, in the Cinnamomeae with a DD formula, offering an exception) are characteristic of the taxonomic sections whose species most resemble each other with respect to chromosome number, as well as septet formulae. This would suggest that in sections as the Gallicanae and the Synstylae that the species are oil the whole not only more closely related to each other than in the sections where more variation occurs, but also that this closer relationship is due to a genetic condition, possibly greater genic homozygosity, which may permit of greater chance for crossing compatibilities and fertility. While some species both within and without these sections may be pointed to as exceptions they do not hide the tendency for sections with least variation in chromosome numbers and septets to show a general effect of this in the crossing relationships of included species. Further studies should throw additional light on this point.

SUMMARY AND CONCLUSIONS

Results of 6,008 pollinations including 118 different attempted species crosses are reported. The data are considered here when the parental species were grouped according to (a) chromosome numbers, (b) taxonomic sections, and (c) genetic groups (Hurst's). Most comparisons were macle on group bases as being more productive of useful conclusions.

Pentaploids, as a group, were the most fertile seed parents, followed by tetraploids and diploids, which were of about equal worth. No effect of apomixis on results from our pentaploid crosses has been revealed to date. Value of a seed parent was but little affected by having its chromosome number equal, larger, or smaller than that of pollen parents. Tetraploids were somewhat the most valuable pollenizers as a group. The comparatively small number of combinations and pollinations with hexaploids and octoploids suggests these to be excellent pollen but poor seed groups.

As in crosses among cliploids per cents of sets and of hip fill indicated greater probability of successful crosses occurring between species belonging to the same taxonomic section than when the species belong to different sections. Species belonging to the Gallicanae and the Caninae were most productive as seed parents. Forms belonging to sections Carolinae and Cinnamomeae were most effective, as a group, for pollen.

The crosses were examined on the basis of Hurst's septet grouping of species. Seed parents with AA, AACC, and DD septets were fertile groups, with the fertility of AA septets usually being increased by the presence of CC. In general similar septets in seed and pollen groups did not increase crossing fertility over cases of combining completely dissimilar septets. There is a tendency for the most fertile septet (seed) groups to fall in the taxonomic sections showing least cytogenetic variation.

LITERATURE CITED

  1. BLACKBURN, K. Chromosomes and classification in the genus Rosa. Amer. Nat. 59: 200-205. 1925.
  2. ERLANSON, E. W. Cytological conditions and evidences for hybridity in North American Wild roses. Bot. Gaz. 87: 443-506. 1929.
  3. GAISER, L. O. (Chromosome number lists). Genetica 8: 401-484. 1926; ibid. 12: 159-256. 1930; Biblio. Genetica 6: 171-466. 1930; ibid. 10: 105-250. 1933.
  4. HURST, C. C. Chromosomes and characters and their significance in the origin of species. Experiments in Genetics. Cambridge, pp. 534-550. 1925.
  5. ———— Differential polyploidy in the genus Rosa L. Verh. V. Intern. Kongr. Vererb. Berlin,. 2: 866-906. 1928.
  6. ———— Embryo-sac formation in diploid and polyploid species of Roseae. Proc. Roy. Soc. London 5cr. B. 109: 126-148. 1932.
  7. RATSEK, J. C., YARNELL, S. H., and FLORY, JR., W. S. Crossing relations of some diploid species of roses. Proc. Amer. Soc. Hort. Sci. 37: 983-992. 1940.
  8. REHDER, A. Manual of Cultivated Trees and Shrubs. Revised Ed. Macmillan Co., New York. 1940.
  9. TACKHOLM, G. On the cytology of the genus Rosa. Svensk. Bot. Tids. 14:300-311. 1920.
  10. ———— Zytologische Studien über die Gattung Rosa. Acta Hort. Bergiani 7:97-381. 1922.
  11. THOMPSON, W. P. Causes of difference in success of reciprocal interspecific crosses. Amer. Nat. 64: 407-421. 1930.
  12. TISCHLER, G. Pflanzliche Chromosomenzahlen. Tab. Biol. 7:109-226. 1931; 11:281-304. 1935; 12:57-115. 1936.
  13. WILLMOTT, ELLEN. The Genus Rosa. 2 Vols. John Murray, London. 1914.