Appendix: Masturbation rates across species

Gory details

• On units
  

  On the question of units, I'd really like to present data in the form of number of masturbation events per day (or some other unit of time). The trouble is that most of the data I have reports only event counts, or rates over hours, during which the animals were observed, and the animals were only observed during times of day when they tended to be awake, so naively converting to days would inflate the rate (animals rarely masturbate while asleep). I'll then convert hours to days in a ratio of 16 rather than 24, on the assumption that animals sleep 8 hours a day (which is clearly wrong in general, but hopefully close enough).

• Nonhuman primates
  
  • Wallis (1983) observed gray-cheeked mangabeys. Masturbation (albeit usually without ejaculation) was common among males. Unfortunately, the Springer PDF is missing too much of the page for me to make out much else.
    
  • Starin (2004) observed western red colobus for a total of 8950 h and saw one male masturbate 3 times and two masturbate once apiece. That means a rate of at most one event per 8950 h / 3, which is at least 4 months.
    

    Among females, one was seen masturbating "at least 35 times", one 6 times, and one just 1 time. There were at least 14 unique females. 8950 h / 6 is at least 2 months.

  • Japanese macaques
    
    • Hanby, Robertson, and Phoenix (1971) observed 69 Japanese macaques for 5 months for at least 3 h a day. They saw "27 cases of male masturbation involving 13 different males". There were 26 sexually active males.
      

      5 months is 152 days, so that's 152 3-h units, or 28.5 16-h units. That's about a month.

      Since 27 cases were distributed among 13 individuals, the most cases any one individual could have is 27 - 12 = 15. So our maximum frequency is 15 events in 28.5 days, which is a reasonably high maximum, so it's not very informative.

      Perhaps it would be better to compute the maximum possible median. This is obtained when the number of events per sexually active male is

      0 0 0 0 0 0 0 0 0 0 0 0 1 1 2 2 2 2 2 2 2 2 2 2 2 3

      and is equal to 1.

    • Thomsen and Soltis (2004) observed 15 male Japanese macaques and give a graph of hourly rates (p. 5). This use of one-zero sampling is a bit suspect, and you can't really calculate each individual's overall rate.
      

      The height of median bars in figure 1, in pixels from the bottom, are: 15, 65, 36, 77, 73, 73, 81, 73, 126, 155, 168, 158, 130. The 10 tick on the y-axis is at 205. So the values are

      v = c(15, 65, 36, 77, 73, 73, 81, 73, 126, 155, 168, 158, 130)
      v = v / 205 * 10
      v = 16 * v   # Converting from 1 / hour to 1 / 16 h
      round(d = 1, sort(v))
      

      	value
      	11.7 28.1 50.7 57 57 57 60.1 63.2 98.3 101.5 121 123.3 131.1

      Which, thought of as daily rates, are really high, particularly compared to Hanby et al. (1971). I'm inclined to blame the zero-one sampling, which takes into account the duration of episodes whereas I want to count the number of episodes. So let's disregard this study.

  • Wolfheim and Rowell (1972) talk about talapoin monkeys. They describe (p. 23) juvenile females masturbating "occasionally" and say that in males, "Prolonged handling of the erect penis culminating in ejaculation was often observed".
    
  • Saayman (1970) observed chacma baboons for 437 h (= 27.3 16-h units) over 93 days. Among 3+15+8 = 26 non-infant males, "Masturbation by males was observed on 14 occasions (7, 5, and 2 times respectively by adult, subadult, and juvenile males)". This means the most events any one individual could have is 7. This is a floor frequency of 7 times per 27.3 days, which is about once per 4 days, which again isn't much of a floor, so let's think about medians.
    

    The median-maximizing arrangement for the adults is (0, 3, 4), for a median of 3, and the median-minimizing arrangement is (0, 0, 7), for a median of 0. That's a low minimum and a high maximum, so there's not much we can conclude.

    So let's instead collapse across age groups. We have 26 individuals and 14 events. The median-maximizing arrangement is

    0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1

    which has a median of 1. That's 1 event per 27.3 days, which is about once a month, which is indeed low.

  • Mootnick and Baker (1994) observed 52 captive gibbons over 17 years. 9 (5 males, 4 females) masturbated at least once, some many times. That's a median of 0.
    

    "For purposes of perspective, no masturbation was observed in 327 hr of systematic observations during a recent, 7-month study of six Hylobates pairs housed at ICGS." As with chimps, there are suggestions that masturbation is caused primarily by unusual rearing conditions.

  • Chimpanzees
    
    • Nishida (1997) reports basically no masturbation in common chimpanzees (687 h of observation)
      

      Masturbation, or self-stimulation of the genitals, has rarely, if ever, been observed in Mahale chimpanzees. Mature and immature males have occasionally been seen briefly to fumble with their erect penises, but this has never been seen to culminate in ejaculation.

    • Rogers and Davenport (1969) describe "frequent" masturbation, and no copulation, by one of 5 male chimpanzees raised in isolation. Perhaps masturbation is a pathological symptom in chimps.
      
    • Kollar, Beckwith, and Edgerton (1968): 3 of 5 adult male chimps did not copulate, and 2 of those masturbated "frequently". The authors seem to attribute this to being "cage-reared and presumably sex-segregated until maturity".
      
  • Bonobos
    
    • Sannen (2003) (p. 76) observed four female bonobos. Their mean hourly masturbation rates (the mean being across days) were 0, .001, .007, .069. Observations were only during waking hours ("from breakfast until 30 minutes after the evening meal").
      

      Multiplying by 16 to get per 16-h unit, we get

      16 * c(0, .001, .007, .069)
      

      	value
      	0 0.016 0.112 1.104

    • de Waal (1995), in "nearly 300 hours" of observing bonobos, observed 39 (solo) masturbation events: 0 by the one adult male, 23 by two adolescent males, 9 by two adult females, and 7 by "immatures".
      

      So among the sexually mature males, the median could be as low as 0 and as high as 11. Restricting to the two adolescents, the median must be 11.5 per 300 hours, or .613 per 18 hours, or once every 1.6 days.

      As for adult females, integrating this data with Sannen (2003):

      ms = range(sapply(1:8, function(f1)
         {f2 = 9 - f1
          median(c(0, .001, .007, .069, f1/300, f2/300))}))
      ms = 1 / (ms * 16)
      round(ms)
      

      	value
      	12 6

      So the median ranges from once every 12 days to once every 6 days.

• Rats - no data
  

  Orbach (1961) and Kihlström (1966) report that rats frequently spontaneously ejaculate and remove the produced "plugs" by grooming their penises. Orbach, Miller, Billimoria, and Solhkhah (1967) observes that the ejaculation really is spontaneous (i.e., it happens even in rats that are prevented from touching their penises). And I wouldn't call the genital-grooming behavior masturbation unless rats do it when there isn't something to clean up.

• Elephants - no data
  

  Masturbation in male elephants is observed, especially in certain parts of the musth process (by, e.g., Jainudeen, Katongole, & Short, 1972), but no freq.

• Cats - no data
  
• Dogs - no data
  

  Kustritz (2005) observes masturbation, but there's no freq information

• Stichbirds - no data
  

  Ewen and Armstrong (2002) - "On 18 occasions male Stitchbirds… were observed neck-rubbing with foliage". Which doesn't seem like very good evidence of masturbation.

• Bears - brown bears - no data
  

  Ishikawa, Sakamoto, Katagiri, and Takahashi (2003) observed 5 sows. 2, which were mated, were not observed masturbating.

  Allegedly, Figure 2 shows masturbation rates. However, the figures are low-quality in the official PDF, and very odd: for example, daily hormone measures were taken, but there seem to be only 5 data points from June 10th to June 20th. And taken at face value, Figure 2 seems to say that on some days, females C and D masturbated 46 (= 4 / hour * (17:30 - 6:00)) times or more.

  I think I'll ignore this paper.

• Bears - sloth bears - no data
  

  Forthman and Bakeman (1992) use instantaneous sampling, so we only have a percentage of time masturbating, not a count of events.

• Dolphin - bottlenose
  

  McBride and Hebb (1948) say "Among males there is a good deal of masturbation", but no freq.

  McBride and Kritzler (1951) likewise say "Masturbation is frequent, particularly in females".

• Dolphin - baiji
  

  Chen et al. (2002) study a single male. Figure 2 shows large seasonal variation in how often it masturbates by rubbing against a wall, ranging from 0 times per hour to an outlier of 40.

  The earlier publication "Cycles of sexual masturbation behavior of a male baiji, 'Qi Qi', in captivity" (http://ssswxb.ihb.ac.cn/EN/abstract/abstract527.shtml) might be more useful, but electronic text might not exist.

  In Chen et al. (2002), the height of the ventral contact bars in Figure 4, in pixels from the bottom, are:

  2, 0, 53, 34, 129, 41, 0, 0, 1, 0, 0, 2, 27, 97, 50, 30, 29, 9, 9, 0, 2, 0, 3, 61, 15, 37, 55, 64, 52, 42, 23, 0, 23, 2

  The 5 tick on the y-axis is at 177.

  v = c(2, 0, 53, 34, 129, 41, 0, 0, 1, 0, 0, 2, 27, 97, 50, 30, 29, 9, 9, 0, 2, 0, 3, 61, 15, 37, 55, 64, 52, 42, 23, 0, 23, 2)
  median(v / 177 * 5 * 16)
  

  	value
  	8.58757062146893

  8.6 times per day? Wow. Well, it was alone in a tank, so maybe it was bored.

• Bulls
  

  Houpt and Wollney (1989) gives daily masturbation rates of nine Holstein bulls.

  median(c(2.88, 2.16, 1.72, 1.58, 1.2, 1.01, 0.96, 0.86, 0.48))
  

  	value
  	1.2

• Donkeys
  

  McDonnell, Henry, and Bristoif (1991): "Pasture breeding Belgian horses (n = 2) and donkeys (n = 2) were studied by continuous direct observation during daylight hours for periods of 9-18 days.[…] One jack, observed for 107 h over a 16-day period, bred 107 times and masturbated 24 times; the other jack, observed for 216 h over an 18-day period, bred 133 times and masturbated 58 times."

  round(d = 3, median(16 * c(24 / 107, 58 / 216)))
  

  	value
  	3.943

• Horses
  

  Tyler (1972) observes masturbation in both stallions and mares, but there's no frequency information.

  Boyd (1988) saw no masturbation in 27 mares. The observation times of each mare, in hours, was:

  5.75 5.25 5.00 5.00 5.00 5.00 18.00 11.00 7.25 4.25 7.25 4.25 5.25 5.25 5.25 5.25 3.50 30.25 23.50 20.75 5.00 13.75 31.75 33.50 5.75 14.00 14.00 5.00 5.75 5.25 14.00 5.00 14.00 14.00 14.00

  So I'm not sure what there is I can say about mares.

  Boyd (1988) (Table 3) observed 0 masturbation events per hour among 2 stallions, .2 among another 8, and .1 among another 8. Note that Boyd (1988) only observed horses from 8 AM to 6 PM. We will assume that outside this time, they were asleep and therefore didn't masturbate. 8 AM to 6 PM is 10 hours. So the daily rates are 0, .2 * 10 = 2, and .1 * 10 = 1.

  McDonnell et al. (1991) observed a mean of 10.3 masturbation events per 24-hour period among 25 domestic stallions.

  Combining Boyd (1988) and McDonnell et al. (1991),

  v = c(rep(0, 2), rep(.2, 8), rep(.1, 8), rep(10.3, 25))
  c(length(v), median(v))
  

  	value
  	43 10.3

• Birds - boat-tailed grackles
  

  Post (1994) observed just 4 incidents of "redirected mounting" in 600 h of observing male boat-tailed grackles. Each incident involved several repeated mounts on the same object. Each incident involved a different individual.

  600 / 16 = 37.5

• Birds - dovekies
  

  Jakubas and Wojczulanis-Jakubas (2008), p. 664:"There was also observation of an unmarked and unsexed individual which mounted a clump of moss in the prelaying period ten times in 2005 and three times in 2006 (probably of the same one because it was observed in the same place)."

• Humans
  
  • NSSHB
    

    This study is the 2009 National Survey of Sexual Health and Behavior. For men, I use Table 5 of Reece et al. (2010), and for women, I use Table 5 of Herbenick et al. (2010).

    ratecats = c(
        "Not in past year",
        "A few times per year to monthly",
        "A few times per month to weekly",
        "2–3 times per week",
        "≥4 times per week")
    
    agecats = c("18-24", "25-29", "30-39")
    
    m = matrix(nrow = 6, byrow = T, c(
        # gender / age / n / Not in past year / A few times per year to monthly / A few times per month to weekly / 2–3 times per week / ≥4 times per week
        1, 1, 260, .185, .169, .250, .208, .188,
        1, 2, 334, .165, .147, .254, .234, .201,
        1, 3, 393, .201, .188, .270, .206, .135,
        2, 1, 192, .365, .286, .245, .073, .031,
        2, 2, 382, .285, .372, .215, .079, .050,
        2, 3, 413, .370, .305, .220, .090, .015))
    
    df = do.call(rbind, lapply(4 : ncol(m), function(ratecol)
       {rate = ratecol - 3
        do.call(rbind, lapply(1 : nrow(m), function(row)
           {gender = c("Male", "Female")[m[row, 1]]
            age = agecats[m[row, 2]]
            row.total = m[row, 3]
            proportion = m[row, ratecol]
            n = round(row.total * proportion)
            data.frame(gender, age, rate)[rep(1, n),]}))}))
    
    c(
       male.n = with(df, sum(gender == "Male")),
       male.median = ratecats[median(subset(df, gender == "Male")$rate)],
       female.n = with(df, sum(gender == "Female")),
       female.median = ratecats[median(subset(df, gender == "Female")$rate)])
    

    	x
    male.n	987
    male.median	A few times per month to weekly
    female.n	987
    female.median	A few times per year to monthly

  • NHSLS
    

    This data is from the 1992 National Health and Social Life Survey (Laumann, Gagnon, Michael, & Michaels, 1994).

    library(foreign)
    d = read.spss("NHSLS.sav", to.data.frame = T)
    names(d) = tolower(names(d))
    attr(d, "variable.labels") = `names<-`(
        tolower(attr(d, "variable.labels")),
        tolower(names(attr(d, "variable.labels"))))
    d2 = droplevels(subset(d,
        !(gender %in% c("Refusal", "DK", "Missing")) &
        !(mast12a %in% c("Refusal", "DK", "Missing"))))
    

    levels(d2$mast12a)
    

    	x
    1	> once a day
    2	Every day
    3	Sev. times a wk
    4	Once a week
    5	2-3 times a mon
    6	Once a month
    7	Every other mon
    8	3-5 times a yr
    9	1-2 times a yr
    10	0 times a year

    c(
        male = median(as.numeric(d2[d2$gender == "Male", "mast12a"])),
        female = median(as.numeric(d2[d2$gender == "Female", "mast12a"])))
    

    	value
    male	8
    female	10

    Those seem really low. I think I'll go with the NSSHB instead, which has the virtues of being nationally representative and more recent.

References