Agriculture Reference
In-Depth Information
In the analysis of Bruneau et al. (2008), the origin of Papilionoideae is placed later
(45 Ma) than that of Lavin et al. (2005), but as they point out, very few samples
were included. Here the estimates of Lavin et al. (2005) are used. In Papilionoideae,
again there are basal groups that lack the ability to nodulate, most notably the non-
nodulating branch of the Swartzioid group and the branch labelled 'various' that covers
non-nodulating members of tribe Sophoreae and related tribes (Fig. 1.3). Again two
nodulating events are possible. Following the second of these events, all subsequent
branches are nodulated, except in the few cases where nodulating ability has been lost.
The genistoid and dalbergioid nodes are dated at 56.4 and 55.3 Ma, respectively (Lavin
et al., 2005) (Table 3.1). In species that have been studied in tribes Crotalarieae and
Genisteae, nodules are formed following direct epidermal penetration, newly infected
cells divide repeatedly, so that infected tissue contains few or no uninfected cells and
transcellular infection threads are not seen (Sprent, 2007 and references therein; see
also Fig. 5.1 of the present topic). Aeschynomenoid nodules are defined by having
an infection via cracks where lateral or adventitious roots emerge and again, newly
infected cells divide repeatedly, uninfected cells in the infected region are rare or ab-
sent and transcellular infection threads are not seen. The major difference between
these two groups is that the former has indeterminate nodule growth and the latter
determinate growth (Figs. 1.1 and 1.2). These features led Sprent (2007) to hypothe-
sise that the earliest nodules were formed following direct infection through wounds
or epidermis and that infection threads were a later development, necessary before
the familiar root hair infection mechanism could occur and bacteria could be trans-
mitted across cells by transcellular infection threads. There are several cases of extant
legumes where infection threads are formed following a non-hair infection. In the
Mimosoideae, aquatic species of
Neptunia
have a crack infection, very similar to that
found in aeschynomenoid nodules, followed by formation of infection threads (Subba-
Rao et al., 1995); in
Mimosa scabrella
infection may occur between epidermal cells, with
infection threads being formed later (Faria et al., 1988) and subsequent nodule de-
velopment proceeding as in some other species of
Mimosa
with a root hair infection
(Elliott et al, 2007a).
Lonchocarpus muehlbergianus
(tribe Millettieae) was found to lack
root hairs and infection was thought to occur through epidermal cells, with later de-
velopment of transcellular infection threads and formation of infected tissue with both
infected and uninfected cells (Cordeiro et al., 1996). There are also several reports of
legumes that normally have a root hair infection bypassing this process under certain
conditions, for example waterlogging in the case of
Lotus uliginosus
(James & Sprent,
1999).
Sesbania rostrata
may adapt both its infection process and nodule development
(determinate or indeterminate) according to whether it is waterlogged or not, but
in both cases transcellular infection threads are formed during nodule development
(Goormachtig et al., 2004). The versatility of nodulation processes in these legumes
from widely different taxonomic groups underlines the suggestion that the primitive
state is a non-hair infection, which is used as a fallback process under certain forms
of stress, especially waterlogging (Goormachtig et al., 2004). Unfortunately there is no
information on the infection of caesalpinioid legumes by rhizobia. The ideas outlined
above are summarised in Fig. 3.2. Comparing this diagram with the chronology given
in Table 3.1 shows that, in addition to the major groups of legumes that had their
origin around 55 Ma, a further group originated in the next 10 million years. This was
