**4. Discussion**

In previous studies, the composition of each histogenic layer in periclinal chimeras such as the ploidy and the graft chimera was evaluated by morphological examination [11,12,20], histological observation [8–10], componential analysis [8], FCM analysis [12], and molecular biological techniques [7,9,10]. To clarify the ploidy level of each histogenic layer in the 2x+4x ploidy chimera, in the present study, flow cytometric analysis, cell observation, and morphological examination of several tissues and organs were carried out. Additionally, the ploidy chimera was crossed with the diploid kumquat to evaluate its reproductive organs of the ploidy chimera.

L1 is distinguished by the dermal system [3]. Especially, the size and the density of the guard cells were employed as the index to confirm the ploidy level in several plant species [19,20]. Epidermis cell size, guard cell size, and guard cell density of mature leaves in the chimera were mostly equal to those of the diploid control. This result showed that the chimera's ploidy level in L1 was diploid.

In the genus *Citrus* and its related genera, on the other hand, the juice sac is often used to analyze the origin of L1 [8,9,12]. Zhang et al. [8] demonstrated the origin of L1 by investigation of the carotenoid composition of the juice sac in periclinal graft chimera "Zaohang" navel orange composed of Robertson navel orange and Satsuma mandarin. Yasuda et al. [12] easily showed, by FCM analysis of the juice sac, that L1 ploidy level of the ploidy chimera Yubeni consists of diploid and tetraploid cells. In another study, however, Sugawara et al. [9] used RAPD analysis on a periclinal graft chimera composed of "Hamlin" sweet orange and Satsuma mandarin to show that cells derived from L1 and L2 were involved in the development of the juice sac. As reported by Nii and Coombe [21], the juice sac develops from epidermal cells and sub-dermal layers. In the present study, we could detect a single diploid peak just like a previous report by Yasuda et al. [12]. From here onwards, it has been considered that the juice sac in kumquats mainly di fferentiated from L1.

For L2 to di fferentiate into parenchymas and reproductive organs, mesophyll cells, pollen, and seeds are often used for the analysis of its origin [9,11,12]. The pollen of the chimera in the present study had as much size and fertility as the tetraploid and the Yubeni, and tetraploid and triploid progenies appeared in the cross between the chimera and diploid kumquat. Furthermore, the sizes of the subepidermal cells in the leaves of the chimera were significantly larger than those of the diploid, and were similar to those of the tetraploid and the Yubeni. For these reasons, it was considered that the L2 of the chimera was tetraploid.

Regarding L3, which is di fferentiated into cambium and pith, Yasuda et al. [12] presumed the ploidy level of L3 in the Yubeni by FCM analysis of the midrib. The FCM analysis of the midrib of the chimera showed tetraploidy. In the cell observations of the vessels and sieve tubes using histological techniques, furthermore, the cells of the chimera were the same size as the tetraploid and the Yubeni. Definitively, it was shown that the polyploidy of L3 in the chimera was tetraploid.

The chimera used in the present study was supposed to be a ploidy periclinal chimera, with diploids in the outermost layer (L1) and tetraploids in the inner layers (L2 and L3) of the shoot apical meristem. The morphological characteristics of the chimera were similar to that of the tetraploid Meiwa kumquat as previously reported [14,22,23]. Especially, it was reported that the fruit of these tetraploid kumquats had a thicker pericarp and a higher sugar content than the fruit of the diploid ones. The albedo, which is the main edible part of kumquats, di fferentiates from L2 [3], so Yasuda et al. [12] showed that the tetraploidization of L2 added to the superior fruit quality of the ploidy periclinal chimera Yubeni (diploids in L1 and tetraploids in L2 and L3). Because the chimera in the present study also had desirable fruit traits for kumquats, such as a thick pericarp, a high sugar content, and a small number of developed seeds, ploidy periclinal chimeras with tetraploids in L2 are very useful for kumquats. Furthermore, in the present study, triploid progenies were obtained from reciprocal crosses with the diploid kumquat. This result indicates that ploidy periclinal chimeras with tetraploids in L2 can be useful as parents for triploid breeding, where seedless fruits can be expected.

In conclusion, the 2x+4x ploidy chimera was confirmed as a ploidy periclinal chimera with diploids in L1 and tetraploids in L2 and L3. We plan to carry out the research not only on the

utilization of the 2x+4x ploidy chimera for triploid breeding but also on the commercial growing of its chimera. Additionally, we need to develop an e fficient production technique for the 2x–4x–4x ploidy chimera type.

**Author Contributions:** M.Y., A.T., H.M., K.Y., and H.K. conceived and supervised research. T.N., M.S., and M.Y. designed experiments. T.N. conducted most experiments. M.S. and T.O. assisted and performed some experiments. M.S. analyzed data and wrote the manuscript. M.Y., A.T., H.M., K.Y., and H.K. supervised the preparation of the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.
