Plant Cell Reports (1998) 17: 685 – 692
K. Padmanabhan · D. J. Cantliffe · R. C. Harrell
D. B. McConnell
A comparison of shoot-forming and non-shoot-forming somatic embryos
of sweet potato [Ipomoea batatas (L.) Lam.] using computer vision
and histological analyses
Received: 27 January 1997 / Revision received: 28 January 1998 / Accepted: 12 February 1998
Abstract Diagnostic structural features for competence
to form shoots were tested among sweet potato embryos
by combining morphological image capture (using a
computer vision system) with anatomical analyses (using
light microscopy). Five major morphological variants
(‘perfect’, ‘near perfect’, ‘limited/no meristematic activ-
ity’, ‘disrupted internal anatomy’, and ‘proliferating’)
were identified among torpedo- and cotyledonary-stage
embryos. Among these, only the first two were found to
be competent for conversion into plantlets. Lack of orga-
nized shoot development in somatic embryos of sweet po-
tato was associated with the following abnormalities: lack
of an organized apical meristem, sparcity of dividing cells
in the apical region, flattened apical meristem, and multi-
ple meristemoids and/or diffuse meristematic activity
throughout the embryo. Diagnostic separation of most
shoot-forming and non-shoot-forming torpedo and coty-
ledonary embryo variants was achieved.
Key words Somatic embryo · Ipomoea batatas ·
Computer vision · Histology
Abbreviations 2,4-D 2,4-Dichlorophenoxyacetic acid ·
IAA Indoleacetic acid · MS medium Murashige and Skoog
medium
Communicated by I. K. Vasil
K. Padmanabhan · D. J. Cantliffe (½)1
Department of Horticultural Sciences, University of Florida,
P. O. Box 110690, Gainesville, FL, 32611, USA
Fax: +1-352-392-6479
R. C. Harrell
704 NE Sixth Street, Gainesville, FL 32601, USA
D. B. McConnell
Department of Environmental Horticulture, University of Florida,
Gainesville, FL 32611, USA
Present address:
1
8633 Datapoint Drive, No 211, Phase II, San Antonio, TX 78229,
USA
© Springer-Verlag 1998
Introduction
Somatic embryogenesis has been induced in many impor-
tant crop and horticultural species (Thorpe 1995). Most re-
search on somatic embryogenesis has concentrated on
developmental embryogenesis, except for a few studies re-
lated to plasticity in morphogenetic patterns of somatic em-
bryos. A variety of somatic embryos develop in culture and
vary greatly in their developmental patterns and their abil-
ity to convert to plantlets. The torpedo-, cotyledonary- and
elongated-torpedo-stage somatic embryos of sweet potato
can convert into plantlets (Chee and Cantliffe 1988; Schul-
theis et al. 1990). However, it is also known that many well-
formed somatic embryos fail to form plantlets (Kerns et al.
1986; Merkle et al. 1990). Therefore, morphological var-
iants exist even among the torpedo- and cotyledonary-stage
embryos that are mature enough to give rise to plants (Chee
and Cantliffe 1988). A comparison of the external mor-
phology with the internal anatomy of somatic embryos may
thus provide valuable insights into their developmental ca-
pabilities, and help identify the variants that can form
plantlets.
A machine vision system can be a valuable tool for iden-
tifying and classifying biological materials (Harrell et al.
1992). The purpose of the present study was to follow the
developmental patterns of somatic embryos via analysis of
external morphology (captured via machine vision) and
internal anatomy in order to identify diagnostic features of
shoot-forming and non-shoot-forming variants. Torpedo-
and cotyledonary-stage sweet potato somatic embryos
were used as test materials.
Materials and methods
Embryogenic culture initiation and maintenanc
Apical and axial shoot tips of sweet potato were used to initiate
embryogenic cultures (Liu and Cantliffe 1984). Embryogenic callus
was gently broken with a sterile glass slide tip and passed through
686
710-µm and 355-µm sieves. The callus collected on each sieve was
rinsed with 3% (wt/vol) sucrose solution. The resultant 355–710 µm
fraction was plated onto callus proliferation medium containing
modified Murashige and Skoog (1962) (MS), with 30 basal medium,
mM KCl, 10 µM 2,4-dichlorophenoxyacetic acid (2,4D) and 1 µM
benzyladenine and 0.7% Phytagar (Gibco) (Chee and Cantliffe
1992). The cultures were maintained in the dark at 30°C in 8-week
cycles.
Embryo production and harvest
Torpedo- and cotyledonary-stage somatic embryos were harvested
after 7 weeks in callus proliferation medium, and transferred indi-
vidually onto small Petri dishes (60×15 mm) containing embryo con-
version medium (MS basal medium with 1.6% sucrose and no growth
regulators; Chee et al. 1992) for root/shoot conversions.
Computer vision and histological analyses
Harvested somatic embryos were subjected to computer vision anal-
ysis (Harrell et al. 1992) on the day of harvest (day 0). The comput-
er vision system consisted of a zoom stereomicroscope (Nikon SMZ-
2T), a color video camera (Pulnix TMC-54G), a 32-bit computer (Mi-
zar 7120) equipped with three image acquisition cards (Datacube
VVG128), and a custom Stoboscopic illuminator (Harrell et al. 1992,
1993). Red-green-blue video signals generated by the camera at a
5-bit resolution were digitized using the three image acquisition cards
(Harrell et al. 1992). A non-blurred image of each embryo was creat-
ed by synchronizing the image acquisition with the flash of the strob-
oscopic illuminator (Harrell et al. 1992, 1993, 1994). Images of 378
embryos were obtained from three individual experiments. Of these,
75 embryos were subjected to histological analyses. Five embryos
were removed from culture at random after 1, 5, 10, 15, and 20 days
and fixed in FAA, dehydrated with a tertiary-butanol series, infiltrat-
ed and embedded with paraplast, sectioned at 10 µm, stained with
toluidine blue and mounted in permount (Jensen 1962). Serial sec-
tions of the embryos were studied and photographed using a Nikon
compound microscope. Sectioned zygotic embryos excised from
Ipomoea batatas seeds (Chee and Cantliffe 1988) were used for com-
parison with somatic embryos.
Results
Histological studies
Longitudinal sections of embryos 1, 5, 10, 15, and 20 days
after transfer to conversion medium were compared with
their respective external morphological images stored in
the computer (vision system). This comparison of exter-
nal morphology and internal anatomy of the embryos
identified five morphological classes among the torpedo-
and cotyledonary-stage embryos: (a) ‘perfect’ (Fig. 1),
(b) ‘near perfect’ (Fig. 2), (c) ‘limited/no apical meristem
activity’ (Fig. 3), (d) ‘disrupted internal anatomy’
(Fig. 4), and (e) ‘proliferating class’ (Fig. 5). Of these
five classes, ‘perfect’ and ‘near perfect’ embryos consti-
tuted about 40% of the total population. The ‘limited/no
apical meristem activity’, ‘disrupted internal anatomy’
and ‘proliferating’ embryo classes constituted the
remaining 60%, with the majority of embryos (about
34%) categorized under ‘limited/no apical meristem ac-
tivity’.
The ‘perfect’ class comprised mostly cotyledonary-
stage embryos, which were either narrow and thin, or wide
and thick, with one to two lobed and sometimes fused cot-
yledons (Fig. 1A,C,E). Cotyledons comparable to those of
zygotic embryos (long, bilobed and folded like wings;
Fig. 7A) were not found. Cotyledonary-stage somatic
embryos were characterized by a vase shape (Chee and
Cantliffe 1988) and possessed a constriction at the shoot
pole.
‘Perfect’ class embryos contained highly cytoplasmic
cells in the shoot apical notch and well-developed vascu-
lar tissue (Fig. 1B,D,F), much like the zygotic embryos
(Fig. 7B). Meristematic cells in the shoot apical region of
embryos form a dome before the initiation of leaf primo-
dia (Nickle and Yeung 1993). Cotyledonary-stage somatic
embryos of sweet potato possessed a dome-shaped shoot
apical meristem after only 1 day in conversion medium
(Fig. 1F). Continued development in conversion medium
(5 days) resulted in formation of leaf primodia (Fig. 1B).
Most embryos of all types had a well-developed root mer-
istem and gave rise to roots.
The ‘near perfect’ class comprised mostly torpedo- to
late-torpedo-stage embryos. The torpedo embryos were
usually long and thin, and creamish white in color with ru-
dimentary or fused cotyledons, and a distinct constriction
at the shoot apex (Fig. 2A,C). Longitudinal sections of
these embryos showed at least a bilayer of densely cyto-
plasmic cells in the shoot apical notch (Fig. 2B) after day
1 in culture. These cells divided and formed a meristematic
dome after 5–10 days (Fig. 2D) and gave rise to a shoot
within a total of 10–15 days. These embryos had well-de-
veloped vascular tissue and root meristem.
The ‘limited/no apical meristem activity’ class was
comprised of mainly torpedo- to late-torpedo-stage em-
bryos. They were usually long and thin, and inverted cone-
shaped with a closed apical end that was either smooth or
rounded without any constriction or depression. These em-
bryos possessed rudimentary or no cotyledons (Fig. 3A,C).
Some of the embryos were vase-shaped with a very thin
and long neck-like hypocotyl. Cells in the shoot apex of
these embryos were vacuolated (Fig. 3B). Even after 20
days in culture, when most embryos of competent classes
Fig. 1A–F The ‘perfect’ class of somatic embryos of sweet potato
was identified by combining external morphological images of 75
embryos with their internal anatomy. The external morphology of
somatic embryos was captured via computer vision analysis on the
day of harvest. Five embryos were selected at random after 1, 5, 10,
15, and 20 days in culture and then subjected to histological analy-
ses in three replications (white bar 1 mm, black bar 0.1 mm). A, C,
E External morphology of ‘perfect’ class embryos with well-devel-
oped cotyledons that are either single or bilobed (arrowheads). B
Light micrograph of the internal anatomy of a ‘perfect’ class somat-
ic embryo (A) after 5 days in conversion medium with a well-devel-
oped shoot apical meristem (arrow). D Light micrograph of internal
anatomy of a ‘perfect’ class somatic embryo (C) after 10 days in con-
version medium with a developing leaf primordium (arrow). F Light
micrograph of the internal anatomy of a ‘perfect’ class somatic em-
bryo (E) at day 1 in conversion medium with a well-formed shoot
apical meristematic dome (arrow)
687
688
Fig. 2A–D The ‘near perfect’ class of somatic embryos of sweet
potato was identified as described for the ‘perfect’ class embryos
(see legend to Fig. 1) (white bars 1 mm, black bars 0.1 mm). A, C
External morphology of ‘near perfect’ class embryos (a-arrowhead,
distinctive constriction at the shoot apical notch, c-arrowhead, cot-
yledons that are either single or fused). B Light micrograph of inter-
nal anatomy of ‘near perfect’ class somatic embryo (A) at day 1 in
the conversion medium with development of at least a bilayer of me-
ristematic cells (arrow). D Light micrograph of internal anatomy of
‘near perfect’ class somatic embryo (C) after 5 days in conversion
medium with a well-developed shoot apical meristematic dome
(arrow)
converted to plantlets, there were no signs of any meriste-
matic activity at the apex (Fig. 3D).
The ‘disrupted internal anatomy’ class was character-
ized by torpedo-stage embryos, which were usually small
and kidney-shaped, and yellowish with some gray dis-
coloration in the interior of the embryo (Fig. 4A,C). These
embryos were typified by the presence of a laterally dis-
placed apical notch with single rudimentary or no cotyle-
dons. There was good vascular tissue development after
day 1 in culture, but no bilayer of meristematic cells
(Fig. 4B). With continued growth in the conversion me-
dium, vascular tissue was completely disrupted with pre-
mature differentiation and distortion of the cell lineage, es-
pecially in the cortex, after day 15 in culture (Fig. 4D).
There was recapitulation of some meristematic activity af-
ter 15 days in culture in the laterally shifted apical notch
in some embryos (Fig. 4E). However, no visible shoot de-
velopment occurred in these embryos even when they were
allowed to develop for a longer time in conversion me-
dium. The embryos started to turn brown externally in
about 15 days and eventually ceased to grow.
Late-torpedo- and cotyledonary-stage embryos consti-
tuted the ‘proliferating class’, and exhibited a definite
smooth curvature in the expanded hypocotyl region
(Fig. 5A,C). Some of the embryos were curved on both
sides of the hypocotyl like a dumbbell. The cotyledons
were either single- or two-lobed, or sometimes fused
(Fig. 5A,C).

Fig. 3A–D The ‘limited/no apical meristem activity’ class of so-
matic embryos of sweet potato was identified as described for the
‘perfect class’ embryos (see legend to Fig. 1) (white bars 1 mm, black
bars 0.1 mm). A, C External morphology of ‘limited/no apical mer-
istem activity’ class embryos (a-arrowhead, no distinctive constric-
tion at the shoot apex, c-arrowhead, no or rudimentary cotyledons
that are single lobed or fused). B Light micrograph of internal anat-
omy of ‘limited/no apical meristem activity’ class embryo (A) at day
1 in the conversion medium with no shoot apical meristem activity
(arrow). D Light micrograph of internal anatomy of ‘limited/no ap-
ical meristem activity’ class embryo (C) after 20 days in conversion
medium with no shoot apical meristem activity (arrow)
Longitudinal sections revealed multiple meristematic
regions throughout the embryos after 5 days (Fig. 5B), and
at the shoot apex after 15 days (Fig. 5D), resulting in cal-
lus formation. Along with callus proliferation either at the
root pole or both the root and shoot pole, some of the em-
bryos gave rise to multiple organized shoots. However, the
embryos retained a well-developed vascular tissue and root
meristem.
689
Discussion
Stuart et al. (1985) suggested that morphological features
are identifiable markers of conversion-competent em-
bryos. Embryo length has been positively correlated with
germination and/or plant formation in zygotic and somatic
embryos (Gray and Purohit 1991; Stuart et al. 1985). How-
ever, heterogeneity, even among torpedo- and cotyledon-
stage somatic embryos, affects the reliability with which
conversion-competent embryos can be manually separated
from non-competent embryos. Torpedo and cotyledonary
somatic embryos may have a normal appearance, but may
be internally abnormal in terms of cellular and tissue dif-
ferentiation (Ammirato 1985). This further complicates the
manual sorting of embryos.
We have identified at least five patterns of development
leading to competent or non-competent somatic embryos
of sweet potato. Lack of conversion in somatic embryos
was traced to anomalies in the shoot apical meristem that
start during the torpedo stage (Ammirato 1985; Merkle
et al. 1990). These anomalies include failure to establish a
690
Fig. 4A–E The ‘disrupted internal anatomy’ class of somatic em-
bryos of sweet potato was identified as described for the ‘perfect’
class embryos (see legend to Fig. 1) (white bar 1 mm, black bar
0.1 mm). A, C External morphology of ‘disrupted internal anatomy’
class embryos with a laterally shifted shoot apical notch (arrowhead).
B Light micrograph of internal anatomy of ‘disrupted internal anat-
omy’ class somatic embryo (A) at day 1 in conversion medium with
no shoot apical meristem activity (a-arrow). D, E Light micrograph
of internal anatomy of ‘disrupted internal anatomy’ class somatic
embryo (C) after 15 days in conversion medium with resumption of
some shoot apical meristematic activity (b-arrow), and disruption of
vascular tissue (c-arrow)
functional shoot meristem (Nickle and Yeung 1993), pre-
mature conversion, extended cell divisions in the shoot
apex (Ammirato 1985), precocious vacuolation in the
shoot cortex, or aberrant cotyledons (Fujii et al. 1990;
Merkle et al. 1990; Quinn et al. 1989). Nickle and Yeung
(1993, 1994) found that loss of conversion capacity in car-
rot somatic embryos was due to abnormal development of
the shoot apical meristem followed by abortion of the mer-
istem.
For all morphological somatic embryo variants in sweet
potato, shoot formation was much less frequent than root
formation, which led to low plant recovery. The lack of or-
ganized shoot development can be attributed to the follow-
ing abnormalities: lack of an organized apical meristem
(Fig. 3B), sparcity of dividing cells in the apical region
(Figs. 3D, 4A,D), flattened apical meristem (Fig. 5B), and
multiple meristemoids and/or diffuse meristematic activ-
ity throughout the embryo (Fig. 5B,D). Of the four abnor-
malities in shoot formation, the first two categories con-
stituted about 80% of the embryos, while the third and
fourth categories constituted about 20% of the embryos.
Auxin appears to play a role not only in those sequences
that precede embryo formation, but also in subsequent mor-
phogenic events in embryo development (Michalczuk
et al. 1992 a,b; Zimmerman 1993). Exogenously applied
synthetic auxin such as 2,4-D stimulates the accumulation
of endogenous indole acetic acid (IAA) (Michalczuk et al.
1992b). This maintains the proliferative state of the callus
and prevents subsequent embryo formation (Michalczuk
et al. 1992b). Transferring the tissue to a 2,4-D-free me-
dium results in a decline in total IAA, to levels low enough
to set up an internal gradient for initiation and maintenance
of polarized growth and subsequent embryo development
(Michalczuk et al. 1992a).
A low conversion rate of somatic embryos may limit
their use in synthetic seed technology. Evidence presented
here and in other studies points towards abnormalities in

Fig. 5A–D The ‘proliferating’ class of somatic embryos of sweet
potato was identified as described for the ‘perfect’ class embryos
(see legend to Fig. 1) (white bars 1 mm, black bars 0.1 mm). A, C
External morphology of embryos with a typical curvature in the ex-
panded hypocotyl (arrowhead). B Light micrograph of internal anat-
omy of embryo (A) after 5 days in conversion medium with flattened
shoot apical meristem (a-arrow). D Light micrograph of internal
anatomy of embryo (C) after 10 days in conversion medium with
multiple meristematic activity at the shoot apex (b-arrow)
the shoot apical meristem as the main cause for a low con-
version rate. Identification of conversion-competent em-
bryos is essential for improved regeneration efficiency, re-
gardless of the hormones used to initiate shoot meristems.
Acknowledgements This is Florida Experiment Station Journal
Series Number R-05217.
References
Ammirato PV (1985) Patterns of development in culture. In: Henke
RR, Hughes KW, Constantin MJ, Hollaender A, Wilson CM
691
(eds) Tissue culture in forestry and agriculture. Plenum, New
York, pp 9–29
Chee RP, Cantliffe DJ (1988) Somatic embryony patterns and plant
regeneration in Ipomoea batatas Poir. In Vitro Cell Dev Biol
24:955–958
Chee RP, Cantliffe DJ (1992) Improved production procedures for
somatic embryos of sweet potato for a synthetic seed system.
Hortscience 27:1314–1316
Chee RP, Schultheis JP, Cantliffe DJ (1992) Micropropagation of
sweet potato (Ipomoea batatas L.). In: Bajaj YPS (ed) Biotech-
nology in agriculture and forestry, vol. 19. High tech and micro-
propagation, 3rd edn. Springer, Berlin, Heidelberg, New York,
pp 107–117
Fujii JAA, Slade D, Olsen R, Ruzin SE, Redenbaugh K (1990)
Alfalfa somatic embryo maturation and conversion to plants.
Plant Sci 72:93–100
Gray DJ, Purohit A (1991) Somatic embryogenesis and development
of synthetic seed technology. Crit Rev Plant Sci 10:33–61
Harrell RC, Bieniek M, Cantliffe DJ (1992) Noninvasive evaluation
of somatic embryogenesis. Biotech Bioeng 39:378–383
Harrell RC, Hood CF, Molto E, Munilla R, Bieniek M, Cantliffe DJ
(1993) Machine vision based analysis and harvest of somatic em-
bryos. Comp Electron Agric 9:13–23
Harrell RC, Bieniek ME, Hood CC, Munilla R, Cantliffe DJ (1994)
Automated in vitro harvest of somatic embryos. Plant Cell Tis-
sue Organ Cult 39:171–183
Jensen WA (1962) Botanical histochemistry – principles and prac-
tice. Freeman, San Francisco
692
Kerns HR, Barwale UB, Meyer MM Jr, Widholm JM (1986) Corre-
lation of cotyledonary node shoot proliferation and somatic em-
bryoid development in suspension cultures of soybean (Glycine
max L. Merr.). Plant Cell Rep 5:140–143
Liu JR, Cantliffe DJ (1984) Somatic embryogenesis and plant regen-
eration in tissue cultures of sweet potato (Ipomoea batates Poir.).
Plant Cell Rep 3:112–115
Merkle SA, Wiecko AT, Sotak RJ, Sommer HE (1990) Maturation
and conversion of Liriodendron tulipifera somatic embryos. In
Vitro Cell Dev Biol 26:1086–1093
Michalczuk L, Cooke TJ, Cohen JD (1992a) Auxin levels at differ-
ent stages of carrot embryogenesis. Phytochemistry 31:1097–
1103
Michalczuk L, Ribnicky DM, Cooke TJ, Cohen JD (1992b) Regula-
tion of indole-3-acetic acid biosynthetic pathways in carrot cell
cultures. Plant Physiol 100:1346–1353
Murashige T, Skoog F (1962) A revised media for rapid growth and
bioassays with tobacco tissue cultures. Physiol Plant 15:473–497
Nickle TC, Yeung EC (1993) Failure to establish a functional shoot
meristem may be cause of conversion failure in somatic embryos
of Daucus carota (Apiaceae). Am J Bot 80:1284–1291
Nickle TC, Yeung EC (1994) Further evidence of a role for abscis-
ic acid in conversion of somatic embryos of Daucus carota. In
Vitro Cell Dev Biol 30:96–103
Quinn J, Simon JE, Janick J (1989) Histology of zygotic and somat-
ic embryogenesis in borage. J Am Soc Hort Sci 114:516–520
Schultheis JR, Cantliffe DJ, Chee RP (1990) Optimizing sweet po-
tato [Ipomoea batates (c.) Lam] root and plantlet formation by
selection of proper embryo developmental stages, size and gel
type for fluidized sowing. Plant Cell Rep 9:356–359
Stuart DA, Nelson J, Strickland SG, Nichol JW (1985) Factors af-
fecting developmental process in alfalfa cell cultures. In: Henke
RR, Hughes KW, Constantin MP, Hollaender A (eds) Tissue cul-
ture in forestry and agriculture. Plenum, New York, pp 59–73
Thorpe TA (1995) In vitro embryogenesis in plants. Kluwer, Dor-
drecht
Zimmerman JL (1993) Somatic embryogenesis: a model for early
development in higher plants. Plant Cell 5:1411–1423