top of page

CRY WOLF Project

Correlating Wolf Behavior with Vocalizations

Wapiti Pack at Elk Creek

Footage Courtesy of Bob Landis

The Holy Grail in bioacoustics is to discover the meanings of various animal sounds.


John and Mary Theberge have been studying wolf vocalizations for over 50 years all over the world. Authors of several books on wolves as well as contributors to a 1971 album produced by Columbia Records and narrated by Robert Redford, titled The Language and Music of Wolves. They graciously contributed their fully-annotated wolf recordings taken in Yellowstone from 2002 through 2018 for use in the Cry Wolf Project. 


Theberge Data Set

Use this link to view the data in full screen.

Screenshot 2024-03-03 103211.png

Coding Guide for Environmental & Social contexts of Recorded Wolf Howls

For more details, refer to: Theberge, J.B. and M.T. Theberge. 2022. "Triggers and consequences of wolf howling in Yellowstone National Park and connection to communication theory". Canadian Journal of Zoology `100: 799-809 (Open Access).

Column 1  “Number” identifies every howl event.  It is also the coupler for the event’s matching sound file where it appears at the end of its identifier clause with an “A” at its end.


Column 2  “Date


Column 3  “Sequence”  Each field trip had identifying events labelled H1, H2, etc.  Links with field notebooks including when it has an “a” at the number, which is continuing howls in the same event.  Otherwise, superseded for purposes of sound analysis by Column 1.


Column 4  “Time”  which sometimes lacks precision with that in the event’s sound file because of extended recordings especially of prolonged events, but the times are close.


Column 5  “Pack” name


Column 6  “Location”  Categories are:



    2=traditional rendezvous site,

    3=traditional resting site,



    See Methods in publication “Triggers and consequences…* for detailed definitions


Column 7  “Location” in territory.  Categories are: C=central, P=peripheral T=trespass.


Column 8  “Howl”  Categories are p=pack, s=single, p/s or s/p=both


Column 9  “If P#”  meaning if howling was by a pack, the number of wolves present


Column 10  “Initiator”  (if known and collared entered is ID number)


Column 11   “All present”  enter y or n


Column 12   “# wolves present”   Redundant with column 9


Column 13  “Others within 2km”  Noted here is the number of wolves in either the same pack (sp) as the sender or in different pack (dp)


Column 14  “Sender activity before and after”  Coded with a slash between before and after.

               -  40 passive – bedded, sitting, walking around, feeding

               -  40a stationary but not visible

               - 18 Play

               - 5  Travel intense

               - 6  Travel slow

               - 6a Milling around

               - or anything else on Howling Data Form Explanation: eating,

                    prey chase, etc.


Column 15  “Sender social situation before/after howl

  • 40 Disturbance by people

  • 41 Disturbance by bear

  • 42 Other disturbance

  • 23 road split

  • 23a excited by scent of foreign wolves

  • 8 – wolf (wolves) coming in to join rest of pack after being away (before it gets there)

  • 9 - Wolf (wolves) howling after reuniting

  • 11 - wolf (wolves) left behind by pack

  • 12 – wolf actively searching for rest of pack (travelling but not just staying where left behind by the pack)

  • 13 – close encounter with foreign wolf(s)

  • 18 – play usually silent

  • 19 during chase or initiation of prey chase usually silent

  • 20 wolf greeting pack member(s) (come together after lying around within 1 km for extended time) and then howl

  • 21 scrum without a howl (silent)

  • 22 wolf or wolves travelling back to a kill after bedding nearby for 1 hr +

  • 24 after a prey chase usually silent

  • 13 close encounter with foreign wolf(s) with or without howl response (redundant and amalgamated with 15 if with howl response). No chase

  • 14 interpack chase with howling

  • 15 interpack howling close  

  • 16 interpack howling distant

  • 16a delayed influence of distant howling pack mates

  • 25 trying to get pack moving, some go, others stay

  • 26 trying to get laggards going in a travelling pack

  • 50 None of list so non triggered

  • 51 not known because not visible.


Column 16  “Receiver SP or DP”  Receiver is the same pack or different pack


Column 17   “Receiver Activity Before/After”   Same choices as in column 14 “Sender Activity Before/After”


Column 18  “Receiver social situation before/after howl”  Same as for sender


Column 19  “Consequence”  Receiver howl and details on receiver/sender travel  coded as follows:   

        HO = Howl response only

        T    = Travel by one or other towards or away from one another only

        HO+T = Wolf(s) do both howl and travel


Then also more detail on travel being:

       Travel “a” = subsequent travel by initial howler(s) to responder;

       Travel “b” = subsequent travel by responding howler(s) to initiator(s).

       Travel “c” = subsequent travel away from one another.


Note: Sometimes but rarely these a, b, or c were used as additional coding in the 2nd consequence. column.


Column 20   “Consequence” being howling interaction between sender and receiver, coded as follows:

  S=single,     A=answered by,     P=pack,     o= means both sender and receiver are in same pack (ie intrapack howling), and no o means both sender and receiver are in different packs (ie interpack howling),  ?=unknown if howling response is same pack or different pack:

     Intrapack                    Interpack                     Unknown     

      SAPo                           SAP                            SAP?

      PASo                           PAS                            SAS?    

      PAPo                           PAP                            PAP?

      SASo                           SAS                            SAS?


In addition in Column 20 was some coding that was rarely used:

C9. Single howls then silent pack comes in

C10. Pack howls then silent single comes in

C11. Single howls then silent single comes in

C12. Pack howls then silent pack comes in


Howl then Travel to Silent Wolf/wolves

C13. Single after howling travels to silent pack

C14. Pack after howling travels to silent single

C15. Single after howling travels to silent single

C16. Pack after howling travels to silent pack


Column 21  “Yipyap”  refers to a pack howl going into a scrum and howling changing to juvenile yipyap, even by adults.  Coded y=yes, n=no


Column 22  “Remarks”  Important description of the event.  Also consistently entered within it, and searchable, is the word “prolonged”, sometimes is capitals and sometimes shortened to “PROLON”.      



The research objective is to identify relationships between previously published environmental/social contexts and consequences of wolf howl events (Theberge, J. M. and M. T. Theberge. 2022. Triggers and consequences of wolf howling in Yellowstone National Park and connection to communication theory. Canadian Journal of Zoology 100: 799-809) and the sound elements and patterns in the recorded howls themselves, ie. the communicative potential to be found in the sound variability.  Communicative potential may refer to emotional/drive states, found to be important in the previous context publication, or possible situation-specific referential communication. 


Explanation of data 

The data are structured around “howl events,” defined in detail in the aforementioned publication but in summary being howls given or received in any amount until broken by 5 minutes of silence.  That arbitrary definition was chosen to accommodate wolves’ most common natural grouping of howls.  The data included here consist of approximately 400 howl events, all with accompanying sound files with an estimated 6 to 10 howls per event to total between 2,400 and 4000 individual howls.  The majority of events however, about 2/3, are pack or chorus howls.  The 1/3 of events by single wolves is about 130 and that time 6 to 10 is close to 800.  One or more single howls do also occur in possible the majority of pack howls, so could add to the total singles, but for our purposes of linking sound features to emotion/drive states would have to be analyzed separately.


Note that for years 2012 and 2013 we entered individual folders rather than files for each event in an effort to group especially prolonged events together but found that not worthwhile so abandoned it thereafter.  Almost all folders for those 2 years just include one howl event, anyhow, and were treated that way in the previously mentioned publication.

Various Considerations

An important question is whether or not to use all howls in all single sender events, or to sample per event.  The issue may be one of data independence. We need statistical advice on this.  Sample sizes would be obviously much more robust with the former. Maybe some sort of Analysis of Variance for within-event versus between-event variance is necessary?

Possibly taking away from analyzable singles might be responses from distant packmates or foreign wolves. Differences in amplitude of howls on the recordings due to distance cannot be depended upon to make this distinction. Whether or not distant wolves responded is coded and easily identified in the Access database (Column 20).

“Prolonged” howl events, again defined in the “context” paper and representing about ¼ of single howl events – but in that they may last over an hour each, contain many individual howls.  As an extreme example, one prolonged howl event in year 2004 (3017) has more than 150 single howls in it.  The problem is that emotions/drives involved may change over such timespans with unnoticed or even noticed changing environmental/social situations, and if such prolonged events were included, they would “muddy” the results. Prolonged events might be interesting however, but should they be kept separate? 

Another consideration in selecting howls for analysis is rejecting ones where the howl is chopped off at beginning, a common occurrence at the beginning of events due to seconds lost setting up recording equipment. The full howl is necessary if length of howl number of pitch breaks or slurs are being noted, etc.

A good multiday sequence to see how a prolonged abandonment of a pup by its pack in its rendezvous site changed its howls is year 2002 H1 to H6.

The previously mentioned event, number 3017 in year 2004, involved 3 wolves for over 2 hours (wolf 255, wolf 379, and an uncollared gray). There was much variation in drops in pitch both between and within groupings, making possibly an opportunity to sort out individual identities.

A special grouping is the 100 howl events where receivers were more than 1 km away were known to b present.   These events would allow examination of howl characteristics for differences in whether sender elicits responses and or travel or is ignored.

Contact Us
bottom of page