Sign in

Luke Rendell

@lrendell.bsky.social
2.3K followers 1K following 234 posts

Reader in Biology @_SMRU_ @SchoolofBiology @univofstandrews. Ocean Yachtmaster for hire! Marine conservation, cetacean & fish behaviour, cultural evolution. He/him. Thalassophile. Sailor. #Fife www.st-andrews.ac.uk/biology/people…

PostsRepliesMedia
Luke Rendell @lrendell.bsky.social · 06/10/2026
Three brilliant people and me ... all the world's a stage, and a beginning 🤣This was a lot of fun, recommend sharing a stage with any of these folks - thanks @nakedprimate.bsky.social for the invite!
060
Luke Rendell @lrendell.bsky.social · 03/09/2026
Such a cool idea - not great for whale language claims however 😏🧪
160
Luke Rendell @lrendell.bsky.social · 27/08/2026
🙌
030
Luke Rendell @lrendell.bsky.social · 22/08/2026
This technique should open opportunities to use existing weakly annotated datasets to train models - but we still need to know and understand more about how models can be constructed to be maximally robust to unseen data. Hint - watch this space for Laia's second PhD paper!
000
Luke Rendell @lrendell.bsky.social · 22/08/2026
The TCNs produced recall rates equivalent to agreement rates between multiple human observers (as close to 'truth' as we will ever get!)
Diagram of the temporal convolutional network (TCN) architecture used to classify recordings based on the presence/absence of sperm whale (Physeter macrocephalus) clicks (left), a temporal block (middle) and a dilated causal convolutional layer with a kernel size k = 3 and a dilation d = 2. The deep stack of dilated convolutions allows the network to capture long-range temporal patterns, making it a suitable architecture to detect the rhythmic vocalisation patterns of sperm whales.
110
Luke Rendell @lrendell.bsky.social · 22/08/2026
Using sperm whale click trains as a testbed, Laia showed how a class of neural network models that take into account temporal structure in detections (recognising that true vocalisations typically come in bouts) - temporal convolutional networks or TCNs - can be trained using such weak annotations
110
Luke Rendell @lrendell.bsky.social · 22/08/2026
but obtaining sufficient detailed auditing and labelling for training models that are robust to unseen data is often impractical, so what is typically available are s0-called 'weak annotations' (e.g., presence/absence of species over several minutes of recording)
100
Luke Rendell @lrendell.bsky.social · 22/08/2026
Passive acoustic monitoring (PAM) can generate huge datasets now, and analysing them for the presence of vocal species is one area where deep learning neural network models (the non-marketing term for the current iteration of AI!) offer real promise ...
110
Luke Rendell @lrendell.bsky.social · 22/08/2026
New paper: Classifying bioacoustic data without individual call annotations using temporal convolutional networks and feature extractors - first (but not last!) PhD paper from @garrobe.bsky.social 🥳 pubs.aip.org/asa/jasa/art... @seamammalresearch.bsky.social 🐋🦑🧪
pubs.aip.org
1126
Luke Rendell @lrendell.bsky.social · 21/08/2026
enhancing our growing awareness of the previously unrecognised flexibility that sperm whales show in the ways they make their living... we also speculate whether this relationship between feeding success and group size differs between the large-scale cultural clans of the females...
020
Luke Rendell @lrendell.bsky.social · 21/08/2026
the reasons why are not known, and are important questions for future research, but clearly there is a balance between the pros and cons of large group formation that is sensitive to local ecology and sperm whales are able to adapt their social structure accordingly...
120
Luke Rendell @lrendell.bsky.social · 21/08/2026
One possibility is that grouping provided foraging benefits. Using feeding success data from Chile and Galápagos we showed that there was a group size that was optimum in terms of individual foraging success, but that was different in the two different places...
A graph of how individual foraging success indexed by defecation rate varies with group size in two locations in the eastern Pacific showing different optimal group sizes
130
Luke Rendell @lrendell.bsky.social · 21/08/2026
The drivers of this variation remain largely unknown however...
A large group of sperm whales, numbering at least 24, photographed from aboveA smaller group of sperm whales, numbering around 10, photographed from above
120
Luke Rendell @lrendell.bsky.social · 21/08/2026
Female sperm whales live in long-term social units that themselves can form larger, temporary groups, but these groups vary considerably in size e.g. between the Pacific and Atlantic: link.springer.com/article/10.1...
link.springer.com
Multilevel Societies of Female Sperm Whales (Physeter macrocephalus) in the Atlantic and Pacific: Why Are They So Different? - International Journal of Primatology
We can examine the evolution of multilevel societies using comparative studies. Intraspecific comparisons are valuable because confounding factors are reduced. Female sperm whales live in multilevel s...
130
Luke Rendell @lrendell.bsky.social · 21/08/2026
New paper! Costs and benefits of group size for sperm whales of the open ocean royalsocietypublishing.org/rsbl/article... w/ Hal Whitehead, Sam Walmslsey and @anaeg.bsky.social 🐋🦑 @seamammalresearch.bsky.social
royalsocietypublishing.org
Costs and benefits of group size for sperm whales of the open ocean
Abstract. Individual feeding success can change with the size of a social group through competition, cooperation, recruitment and other processes. Female s
12419
Luke Rendell @lrendell.bsky.social · 26/07/2026
Exceot they are not vowels or even "vowel-like". In any way. At all. The entire myth is built on fundamental misunderstanding or misrepresentation of sperm whale sound production - which is why the word 'vowel' appears only in the PR and not the actual peer-reviewed paper 🙄
Screenshot of an article in the Guardian about claims regarding vowel sounds im sperm whales highlighting dubious linguistic gymnastics by the study's defenders
053
Luke Rendell @lrendell.bsky.social · 26/07/2026
Funny that the paper itself doesn't contain the word 'vowel" - because unlike the PR it was subject to critical peer review - c'mon David, you're a better scientist than this...
000
Luke Rendell @lrendell.bsky.social · 26/07/2026
So cool! 🌕 🐒
021
Reposted by Luke Rendell
Jasper Blom @jasprblom.bsky.social · 26/07/2026
What a powerful way to illustrate shifting baseline syndrome. And how sad it sounds... www.theguardian.com/environment/...
theguardian.com
Listen to Britain’s dawn chorus of 1976: the dramatic loss of birdsong in 50 years
Guardian recreates audio landscape of past filled by loud morning symphony before 73m wild birds were lost
25232
Luke Rendell @lrendell.bsky.social · 13/07/2026
New chapter: Sperm Whales in Handbook of the Mammals of Europe. Led by Sanna Isojunno, a comprehensive review of knowledge of sperm whales in European waters - happy to see this out! @jlkershaw.bsky.social @seamammalresearch.bsky.social 🐋🦑 link.springer.com/rwe/10.1007/... DM yr email for PDF
link.springer.com
Sperm Whale Physeter macrocephalus Linnaeus, 1758
This comprehensive species-specific chapter covers all aspects of mammalian biology, including taxonomy, global and European distribution, genetics, morphology, life history, ecology, behavior, parasi...
0164
Luke Rendell @lrendell.bsky.social · 09/07/2026
Editor's pic no less! pubs.aip.org/asa/jasa
Homepage of the Journal of the Acoustical Society of America showing our paper as an editor's pick
050
Luke Rendell @lrendell.bsky.social · 09/07/2026
New paper! I am now officially a bowhead song researcher thanks to Seth's cool new paper - see thread for details - there is so much work to do on bowhead song and I am excited to continue this collaboration! @seamammalresearch.bsky.social @taylorhersh.bsky.social @sethharris15.bsky.social 🐋🦑🧪
1177
Luke Rendell @lrendell.bsky.social · 24/06/2026
Thanks Lorraine!
000
Luke Rendell @lrendell.bsky.social · 24/06/2026
More coverage here: bsky.app/profile/theg...
132
Luke Rendell @lrendell.bsky.social · 24/06/2026
So it's sobering to reflect on the span of human evolution that has occured in the same region over the same time span! My Spanish colleague and friend (and co-author) Txema from @asociaciontursiops.bsky.social puts it well...
Dr Txema Brotons, of Asociación Tursiops the Spanish team involved in the study said “this finding reminds us that the cultural history of the Mediterranean does not belong exclusively to humans. While the civilizations of the Mare Nostrum were developing their own languages, customs and identities, sperm whales were also passing down their vocal traditions from one generation to the next. The Mediterranean is therefore a space of shared cultural diversity, where the evolution of human culture and animal culture has coexisted for thousands of years."
141
Luke Rendell @lrendell.bsky.social · 24/06/2026
and it has arisen in the last 20K years, as this is when population genetic studies place the entry of sperm whales to the Med - they all share the same mitochondrial haplotype so it is plausible they are all descended from the same founding matriline onlinelibrary.wiley.com/doi/10.1111/...
onlinelibrary.wiley.com
Genomics reveals the role of admixture in the evolution of structure among sperm whale populations within the Mediterranean Sea
In oceanic ecosystems, the nature of barriers to gene flow and the processes by which populations may become isolated are different from the terrestrial environment, and less well understood. In this...
130
Luke Rendell @lrendell.bsky.social · 24/06/2026
This variation is clearly cultural as there are documented movements between the basins - see: www.sciencedirect.com/science/arti... onlinelibrary.wiley.com/doi/10.1002/...
sciencedirect.com
131
Luke Rendell @lrendell.bsky.social · 24/06/2026
We also mapped all the dialect reports we could find in the literature on these whales and produced a dialect map that shows a clear divide that falls between Italy and Greece..
Comparison of our results (filled circles) with known or probable dialect group detections (colours) in previous Mediterranean sperm whale research (other symbols) [31–36,61]. Note that for other papers, assigned dialect groups represent our best guess based on the available information, but should not be regarded as definitive without further quantitative analyses. See the electronic supplementary material, table S6 for additional details on past work and our rationale for dialect group assignments of previous studies.
121
Luke Rendell @lrendell.bsky.social · 24/06/2026
Moreoever, there was a clear directionality - sometimes, social groups in the east engaged in throwback sessions where they made a lot of the western codas! but we never saw the inverse - so the eastern whales clearly retained the assumed ancestral dialect..
Representation of Mediterranean coda repertoires showing two detected dialect groups and their corresponding identity coda types. Top: average linkage hierarchical clustering dendrogram shows similarity among sperm whale coda repertoires recorded in two locations in the Mediterranean Sea. Repertoires that are more similar in terms of probabilistic coda type usage cluster closer together and have a higher correlation value (y-axis). The geographical source of each repertoire is listed below the dendrogram (red B, Balearic Islands; blue H, Hellenic Trench). This tree was constructed in IDcall using critfact = 15 and minrep = 15 (electronic supplementary material, table S3; see [17] for a full description of IDcall parameters). Middle: heat map depicts probabilistic identity coda type usage (rows) for each repertoire (columns). Arbitrary identity coda type labels are on the left and identity coda type centroid rhythm plots, where each dot represents one click and the coda type is coloured by dialect group, are on the right; 47.9% of the codas used to generate the dendrogram are represented in this heat map. Labels beneath show year of recording. Bottom: stacked bar plots show the posterior probabilities of repertoire assignments to dialect groups. Bottom labels show the social unit identity when known and the number of codas in each repertoire in parentheses. Along the Hellenic Trench, known social units include: Antikythira, Bestend, Casper, Chromo, Falasa, Ippolyti, Keri, Kythira, Lakon, Palaio, Pylos, Veneti, Zak, Zakpylos and Zante. If multiple social units were recorded together, their names are appended (e.g. PylosVeneti). If a social unit was recorded with additional unknown whales, the social unit’s name is appended with UNK (e.g. PylosUNK). Note that social unit was unknown (and is left blank) for the Balearic Islands coda repertoires.
131
Luke Rendell @lrendell.bsky.social · 24/06/2026
This revealed a striking vocal divergence between animals in the east and western basins - the eastern population had evolved a faster version of the western 3+1 coda with a longer last interval (so 3++1) and it met the criteria for being an identity coda that differentiates the two populations..
Graph of inter-click intervals (ICIs) of all four-click codas (n = 3536) recorded off the Balearic Islands (red) and along the Hellenic Trench (blue). Each dot represents the time at which a single click is produced within a four-click coda. Moving from left to right with the first click at time 0, each coda is thus represented as a horizontal sequence of four dots. Codas are ordered by the time between the first and second clicks (i.e. the duration of the first ICI). Note the clear separation by recording location, with longer duration four-click codas recorded in the west (red) and shorter ones recorded in the east (blue). Also note the regional asymmetry, with very few short duration/fast tempo four-click codas produced by whales recorded in the west (i.e. little red within blue) but more long duration/slow tempo four-click codas produced by whales recorded in the east (i.e. relatively more blue within red).
131
Luke Rendell @lrendell.bsky.social · 24/06/2026
Mediterranean coda dialects are fascinating, even if I say so myself! 🤣 Collectively, we gathered recordings from social groups in the western and eastern basins of the Med over a 20 year study period and analysed the coda patterns from >5000 codas collected over 116 days with whales..
A map showing a series of sampling points in red around the Balearic Islands and blue along the Hellenic Trench. Figure caption: Map of coda repertoire recording locations. Bottom left inset shows the Mediterranean Sea, with the relevant area boxed. Each circle denotes one sperm whale repertoire. Filled circles indicate well-sampled repertoires (with ≥25 3–10-click codas) and empty circles indicate small repertoires (with <25 3–10-click codas). Note that all filled circles have the same opacity; seemingly darker circles are spatially proximate repertoires. All repertoires were used for coda type classification but only well-sampled repertoires were used in our hierarchical cluster analysis. Balearic Islands repertoires are in red and Hellenic Trench repertoires are in blue. The dashed line is approximate and indicates the shallow waters between Sicily and Tunisia which broadly separate the western and eastern basins.
141
Luke Rendell @lrendell.bsky.social · 24/06/2026
New paper! Dialect variation in Mediterranean sperm whales shows evidence of cultural evolution in an isolated population tinyurl.com/yrfepekh with @taylorhersh.bsky.social @asociaciontursiops.bsky.social and the Pelagos Cetacean Research Institute 🐋🦑🧪 @seamammalresearch.bsky.social
A group of sperm whale adult and calves at the surface taken from above. Credit Asociación Tursiops
14923
Reposted by Luke Rendell
Tay (Antonio J. Osuna Mascaró) @biotay.bsky.social · 13/06/2026
1/4 The enormous complexity of algae fishing Researchers have studied in detail how chimpanzees in Guinea fish for algae. They prepare their tools (up to 4 m long) in advance, and use 22 different techniques. (paper) onlinelibrary.wiley.com/doi/10.1002/...
13622
Reposted by Luke Rendell
Cat Hobaiter @nakedprimate.bsky.social · 21/04/2026
I like ba-o-babs and I can not lie… but as it turns out I’m not the only one! Fongoli chimpanzees incorporate baobab hollows into their vocal drumming displays!! 🧪🐵🥁 Are you sitting comfortably… 1/n #newpaper #sciencestorytelling brill.com/view/journal...
A black and white photo of a large fat baobab tree trunk with a hollow
2367
Reposted by Luke Rendell
Alexa Robles-Gil @alexaroblesgil.bsky.social · 29/04/2026
Many of us have seen videos of humans swimming with orcas. The reality, however, is much different. Over the course of nearly a year, I traveled to Norway and Mexico to report on the complexity of swimming with these apex predators. Read more in @nytimes.com gift link www.nytimes.com/2026/04/29/s...
nytimes.com
Clear Waters, Murky Morals: When Humans Swim With Killer Whales
2187
Reposted by Luke Rendell
Camille Troisi @camilletroisi.bsky.social · 15/04/2026
📢Fully-funded #PhD opportunity with us: 🐦 Quantifying (social) learning and social behaviour in an urbanised world 🐦 Position at @ceec-research.bsky.social 📆29th May deadline More info: tinyurl.com/yz28s96x Apply: tinyurl.com/2wpkb64y #cognition #socialbehaviour #fieldwork #birds 🙏Please Share
Great tit perched on a branch, with beak slightly open. 
Photo credit: Alizée Vernouillet
28398
Reposted by Luke Rendell
Franca Eichenberger @francae.bsky.social · 27/04/2026
Really nice to see our recent paper picked up in a Dispatch. It highlights a point we care a lot about: even in a conservation success story like humpback whales, recovery is about more than numbers.🐳 Dispatch: www.cell.com/current-biol... @currentbiology.bsky.social @seamammalresearch.bsky.social
cell.com
Marine biology: Don’t call it a humpback comeback!
As a population recovers from extreme exploitation, it is not simply a matter of how many individuals are in the population but who they are, which can affect sexual selection. A long-term study of hu...
1189
Luke Rendell @lrendell.bsky.social · 24/04/2026
So exciting to see sperm whale dialect data being collected from the Turkish eastern Med waters 👏 Aylin Akkaya - and presented with good humour 🤣 sperm whale squeals can indeed sound like drowning cats! watch this space for more developments on this topic! 👌🦑🧪🐋 #ECS2026
Aylin Akkaya in front of a title slide about sperm whales 🥳
0184
Luke Rendell @lrendell.bsky.social · 24/04/2026
@denrisch.bsky.social sent gliders bearing soundtraps around the famous Stanton Bank - dolphins on the bank, porpoises off it...super cool work! #ECS2026 🦑🧪🏴󠁧󠁢󠁳󠁣󠁴󠁿🐋
Denise Risch and a glider PAM methods outline slide
1102
Luke Rendell @lrendell.bsky.social · 24/04/2026
Georgina Whittome with important findings about blue whale distribution and ENSO patterns in the eastern Pacific 👏 @seamammalresearch.bsky.social #ECS2026 #MSc_MMS_St_A 🐋🦑🧪🏴󠁧󠁢󠁳󠁣󠁴󠁿
Georgina and the talk title slide
073
Luke Rendell @lrendell.bsky.social · 24/04/2026
Day 3 #ECS2026 🥳 Super thoughtful and thought provoking plenary on importance of taking perspective in conservation science from Paul Thompson this morning 🐋🦑🧪
Paul Thompson celebrating the growth in marine mammal research in previous decades
040
Luke Rendell @lrendell.bsky.social · 23/04/2026
@oceangill.bsky.social looks beyond Europe to summarise coastal cetacean conservation issues and massive data gaps around the Indian Ocean, focussing on the concerning status of humpback dolphins and establishing a vital international network #ECS2026 🦑🐋🧪 @seamammalresearch.bsky.social
@oceangill.bsky.social in front of her title slide
093
Luke Rendell @lrendell.bsky.social · 23/04/2026
Minke whakes in Western Scotland need dynamic management of protection and entanglement mitigation as distribution changes across seasons awesome work - local fisher knowledge can be vital! Nice work @timawbery.bsky.social #ECS2026 🐋🦑🧪
Tim with a slide of Western Scotland, God's own country 🤣
171
Luke Rendell @lrendell.bsky.social · 23/04/2026
Cognitive control of vasodilation evidenced by near infra red spectroscopy - so cool! 🥳
020
Luke Rendell @lrendell.bsky.social · 23/04/2026
Day 2 #ECS2026 @eva-mariab.bsky.social telling us about diving physiology explored with foraging experiments at @seamammalresearch.bsky.social #MSc_MMS_St_A 🐋🦑🧪
@eva-mariab.bsky.social and their title slide
142
Luke Rendell @lrendell.bsky.social · 22/04/2026
@seamammalresearch.bsky.social
000
Luke Rendell @lrendell.bsky.social · 22/04/2026
So chuffed to see @ellengarland.bsky.social and I's #MSc_MMS_St_A project student @sethharris15.bsky.social presenting his awesome work on bowhead songs at #ECS2026 🥳🐋🧪
Seth and I standing in front of a poster describing bowhead whale song research
092
Luke Rendell @lrendell.bsky.social · 22/04/2026
Full disclosure I co-supervised @amsouth.bsky.social but my unbiased and objective opinion that his work on humpback song rhythms is awesome! 🤣 @seamammalresearch.bsky.social #ECS2026 🐋🦑🧪
Alex South and his title slide about rhythmic variability in humpback song 😎
0112
Luke Rendell @lrendell.bsky.social · 22/04/2026
Great to see #MSc_MMS_St_A graduate Thomas Webber blazing trails to automatically identofying Rissos dolphin clicks - see cool! #ECS2026 🐋🦑🧪
Thomas Webber telling us about Rissos dolphins in Scotland 🏴󠁧󠁢󠁳󠁣󠁴󠁿
031
Luke Rendell @lrendell.bsky.social · 22/04/2026
Marleine Meister mesmerising with bowhead song, showing how vocal behaviour varies with ice cover. Amazing! #ECS2026 🐋🦑🧪
an audience fascinated by a spectrogram of bowhead song
020