Abstract:
Vocal communication is an essential behaviour for many animals, allowing them to recognise,
assess and coordinate with one another. Some social animals also benefit from vocal learning, which
is an ability to develop and shape vocal repertoires based on auditory templates and acoustic
environments. This ability enhances communication by enabling individuals to adapt and fine-tune
their vocal signals. Hummingbirds (order: Apodiformes, family: Trochilidae), parrots (order:
Psittaciformes), and oscine songbirds (order: Passeriformes, suborder: Passeri), were thought to
have each independently evolved this ability, due to their distant phylogenetic relationship. Among
these birds, Passeriformes have the most diverse life history traits and vocal repertoires, providing
us with an excellent opportunity to understand how and why vocal learning evolved. However, the
avian phylogeny has recently been dramatically reassembled; and the new phylogeny has the
potential to transform our current understanding of vocal learning. In the new phylogeny, the New
Zealand wrens (Acanthisitti) are now the oldest passerines, and a crucial link between parrots and
songbirds, which includes both oscines (canonical learners) and suboscines (assumed non-learners).
Thus, a thorough examination of New Zealand wren vocal abilities may provide critical insights into
whether vocal learning is a binary (yes/no) trait and whether this ability actually evolved in the
common ancestor of parrots and songbirds. Here, I undertook this research from four angles.
First, I conducted a global literature review of vocal development studies in birds across the
avian phylogeny and vocal learning abilities. Vocal development is considered a key feature for
classifying taxa in the classical binary division: learners display pronounced developmental stages
and prolonged development, while non-learners display no clear stages and have rapid vocal
development. However, I found that the empirical data on which these classifications are narrowly
focused on a handful of model species, on songs rather than calls, and exclude females almost
entirely. This uneven documentation creates a skewed view of how vocal development patterns
reflect vocal learning, and even among the existing data the patterns are not what the binary
hypothesis would predict. In the remaining chapters, I addressed this gap by assessing the vocal
communication system of one of the two living species of New Zealand wrens, the titipounamu or
rifleman (Acanthisitta chloris granti). I quantified their vocal repertoire and described nine adult call
types and three juvenile call types. I also found contextual variations in the acoustic parameters of
some call types and discuss their social relevance and functions in communication. I then
investigated titipounamu vocal development and found that the onset timings of juvenile call types
follow systematic stages. In addition, the period of transition of one juvenile call to the adult form
extends beyond fledging and dispersal. These patterns indicate that titipounamu vocal development
are more similar to learning oscine songbirds than non-learning species. Finally, I tested for plasticity
in adult vocalisations to determine if birds are adjusting their vocalisations – an ability usually
attributed to learning. I found that some adult contact calls have more similar structures between
partners than strangers, and individuals show differences across years. These results support the idea that these birds are capable of vocal plasticity. This thesis highlights the importance of
fundamental studies on the repertoire, development, and variation of vocalisations of
underrepresented species to further advance our understanding of vocal learning and its many forms.
Taken together, my findings on the protracted vocal development and plastic nature of adult
contact calls indicate that rudimentary vocal learning is likely present in the titipounamu. Given that
New Zealand wrens have a simple syrinx (vocal organ), distinct from that of a vocal learner, these
living fossils offer a new perspective on vocal learning in its rudimentary form. These results
undermine the longstanding paradigm that used a simple binary classification of learner versus non-learner and suggests that rudimentary vocal learning is much more widespread, and evolved much
earlier, than previously thought. Thus, these findings support the new hypotheses arguing for vocal
learning as a continuum, spectrum, or collection of modular abilities.