Math 142 Section 1

Homework 1
Due Friday, October 11

(1) The following questions will give you some practice on how to write word equations.
(a) You are building a mathematical model for the population of kakapo (an endangered
species of flightless parrots that live in New Zealand). These slow-moving birds are
endangered because they are eaten by predators such as rats. Write down a word
equation relating the population Nk , in a specific wildlife reserve, in one year to the
population Nk+1 in the next year. Your equation will include the following terms:
• Number of kakapo born in the wild
• Number of kakapo removed for captive breeding
• Number of kakapo reintroduced into the wild from captive breeding
• Number of kakapo killed by rats
Nk+1 = Nk + number born in wild - number removed for breeding + number reintro-
duced - number killed by rats
(b) Recurrence equation models can be used to model a wider range of phenomena than
just population growth. One application area is pharmacokinetics: the study of how
medications move through the human body. Suppose we are building a model for how
a medication like ibuprofen (“Advil”) passes from a patient’s gut into their blood, and
from there enters their cells or is eliminated from their body by their kidneys. Write
down a word equation for the amount of ibuprofen in the patient’s blood t hours after
they start a course of treatment with ibuprofen. The word equation should include the
following terms:
• Amount of ibuprofen in the patients blood after t − 1 hours
• Amount of ibuprofen absorbed from the gut into the blood
• Amount of ibuprofen absorbed into tissues
• Amount of ibuprofen filtered from blood by kidneys
• Amount of ibuprofen reabsorbed into blood from tissues

Nt = Nt−1 + amount absorbed from gut - amount absorbed into tissues + amount
filtered from blood - amount reabsorbed from tissues

(2) Saving the kakapo. The kakapo is an endangered species of flightless parrots that live in
New Zealand. These slow-moving birds are endangered because they are eaten by predators
(such as rats). You are modeling the size of the population of the kakapo in an island
reserve in New Zealand. You want to use the mathematical model to predict the size of the
population. The data in this question are taken from Elliot et al. (2001).
(a) You start by writing a word equation relating the population size Nk in year k, that
is, k years after the study began, to the population size Nk+1 in the next year:

Nk+1 = Nk + # of birds born in one year − # of birds that die in one year

We will now derive terms for the birth and death rates.
(i) To estimate the number of birds born, assume that half of the birds are female. A
female bird lays one egg every four years. However, because of the large numbers
of predators (mostly rats), only 29% of hatchlings survive the first year. Use this
data to show that the number of births is 0.03625 ·Nk .
1
2

Number of births = (proportion of birds that are female) × (proportion of females

that lay eggs per year) × (hatchling survival rate) × (current population). Using
the data provided, that means number of births is 12 × 14 ×0.29×Nk = 0.03625Nk .
(ii) Kakapo life expectancy is not well understood, but we will assume that they live
around 50 years. That is, in a given year, one in 50 kakapo will die. What is the
corresponding number of deaths?
1
Number of deaths is 50 Nk .
(iii) Assume that the starting population size on this island is 50 birds (i.e., N0 = 50).
Calculate the predicted population size over the next five years (i.e., calculate
N1 , N2 , ..., N5 ).
1
The model is Nk+1 = Nk +0.03625Nk − 50 Nk . Calculating using this model gives:
k 0 1 2 3 4 5
Nk 50 50.8125 51.6382 52.4773 53.3301 54.1967
(iv) When (if ever) will the population size reach 100 birds? What about 200 birds?
(You will find it helpful to derive an explicit formula for the size of the population
Nk ).
The explicit formula is Nk = (1 + 0.03625 − 1/50)k N0 = 1.01625k N0 . This means
that the population size will reach 100 birds when

100 = 1.01625k ∗ 50
⇒ 2 = 1.01625k
⇒ ln 2 = k ln 1.01625
⇒ k = ln 2/ ln 1.01625 ≈ 43 years.

Using a similar argument, the population will reach 200 birds when

k = ln 4/ ln 1.01625 ≈ 86 years.

(b) Using your model from part (a), you want to evaluate the effectiveness of two different
conservation strategies:
(Strategy 1) If the kakapo are given supplementary food, then they will breed more
frequently. If given supplementary food, rather than laying an egg every four years, a
female will lay an egg every three years.
(Strategy 2) By hand-rearing kakapo chicks, it is possible to increase their one-year
survival rate from 29% to 50%.
(i) Write down a recurrence equation for the population size Nk if strategy 1 is
implemented. Assuming N0 = 50, calculate N1 , N2 , ..., N5 .
Nk = Nk + 12 × 13 × 0.29 × Nk − 50
1
Nk ≈ (1.02833)Nk .
k 0 1 2 3 4 5
Nk 50 51.4167 52.8735 54.3716 55.9121 57.4963
(ii) Write down a recurrence equation for the population size Nk if strategy 2 is
implemented. Assuming N0 = 50, calculate N1 , N2 , ..., N5 .
Nk = Nk + 21 × 14 × 0.5 × Nk − 50
1
Nk = (1.0425)Nk .
k 0 1 2 3 4 5
Nk 50 52.1250 54.3403 56.6498 59.0574 61.5673
(iii) Which conservation strategy gives the biggest increase in population size?
 3

The second strategy is better, because the growth rate is larger and so the pop-
ulation will grow faster.
(3) Yeast cell growth. You are trying to model the growth of a population of yeast cells, and
to compare to this model to real data. The data we will use is from a paper by Carlson.
You should start by downloading the data from the BioQuest website:
https://bioquest.org/numberscount/data-details/?product id=31395
The data consist of a count of yeast cells (the count is measured in terms of millions of cells
per mL of the growth medium) as a function of time.
(a) Start by discussing how, if you watched a group of cells for some interval of time, you
could estimate their mortality (death) rate m.
Answers may vary, but one possible strategy: We know that cells reproduce by splitting.
If we chose ∆t to be the amount of time it takes for a yeast cell to go through its cell
cycle, then b = 1, and the model might be N (t + ∆t) = (1 + ∆t(1 − m)N (t). So, by
counting the number of cells at a given time t, and then comparing it to the number
of cells at time ∆t later, we can estimate m.
(b) By plotting the data on a computer, show that the population size in Carlson’s ex-
periment does not grow exponentially. Your plot should include axis labels and an
explanatory title. Why doesn’t the population size grow exponentially?
The curve should look exponential, then level off. Axes should be labeled (but don’t
take off points this time). The population doesn’t grow exponentially because it is
either resource-limited (space, nutrients, etc.) and/or the cells stop dividing when
they become too crowded. In other words, there is a carrying capacity.
(c) In fact, in the first few hours of the experiment, the population growth is approximately
exponential. Explain how to show this from the data.
If the growth is exponential, the growth factor should be constant over a fixed time
period ∆t. In other words, we should have Nk+1 = ρ0 Nk for ρ0 constant, so in the data,
N
we should look for Nk+1 k
approximately constant for over the first few hours (small k)
of the experiment.
(d) Use the experimental data to estimate the growth rate R0 of the yeast cells when the
population is growing exponentially. There may be some different answers here, but one
possibility is: Using the strategy described above, we can estimate the growth factor.
Looking at the ratios of the cell count at every hour, we get: N1 /N0 ≈ 1.9062, N2 /N1 ≈
1.5847, N3 /N2 ≈ 1.6276, N4 /N3 ≈ 1.5063, N5 /N4 ≈ 1.6751. Assuming that the differ-
ence in these growth rates is just due to noise, we could estimate the growth factor by
taking the average to get ρ0 ≈ 1.66, or R0 = ρ0 − 1 = b − m = 0.66.

(4) Growth of the European settler population in the United States. In this exercise,
we will develop a mathematical model for the population of European settlers in the United
States. We will build a mathematical model to predict the year-to-year changes in the Eu-
ropean settler population, starting in 1630, where the population size was 4646 individuals.
Here is some data that will be helpful to you: at the start of our study, life expectancy
in this population was 35. On average, each woman had 6 children. We will start by deter-
mining how to use these data to fit the parameters b (birth rate) and m (mortality rate) in
the following equation:

Nk+1 = (1 + b − m)Nk ,

where k is the number of years since 1630.

4

(a) Birth rate: assume for simplicity’s sake that half of the population is female and that
women in the population have 6 children distributed randomly over 35 years (their life
3
expectancy). Explain how this translates into a per-person yearly birthrate of b = .
35
Per-person yearly birth rate = proportion that is female × proportion of women who
6
give birth in one year = 12 35 3
= 35 . In words: since women are half of the population,
the per-person yearly birthrate is half of the women’s birthrate (because men give birth
to 0 children per year).
(b) Let’s interpret the life expectancy data as saying that a randomly chosen individual
has a likelihood of one-half of reaching 35 years old. Equivalently, if we start with a
large cohort of individuals (and ignore births, to only follow the individuals initially
present), then if the initial population size is N0 , we expect that after 35 years, the
N0
population size will have dropped to . Find the corresponding mortality rate m.
2
After 35 years, the explicit formula for the population where births are ignored will be
N35 = (1 − m)35 N0 ,
so since we know that the population will be cut in half, we get
N0
= (1 − m)35 N0
2
1
⇒ = (1 − m)35
2
 1/35
1
⇒m=1− ≈ 0.02.
2

(c) A lot of assumptions and simplifications go into our model. Describe two main issues
with the simplifications in our model (there are many possible answers here). Discuss
some ideas about how you might change the model to deal with these simplifications.
Answers will vary. Need to put thought into their answer and actually discuss how to
change the model to get the point here.
(d) Compare your model (the original model from parts a and b) with the real US settler
population size data on the Wikipedia page:
https://en.wikipedia.org/wiki/Demographic history of the United States
The data is in the table marked “Historical Population.” Make a plot of your data with
∆t = 1, and compare it with the real data for the years 1630, 1640, 1650, 1660, 1670,
1680 (i.e., plot your predicted population size as a function of time with the real data
on the same axes). Your plot should be made using a computer and should include
axis labels and an explanatory title.
Plot should look reasonable, be clear, and include axis labels and title.
(e) Our model does not include the contribution of immigration and emigration: i.e., of
new settlers arriving in the US or of settlers leaving the US. Suppose that settlers
arrive at a constant rate I0 per year. The recurrence equation governing population
growth now becomes:
Nk+1 = (1 + b − m)Nk + I0 .
Try different values for the rate of immigration: I0 = 200, 500, and 1000. For each
value of I0 , plot the predicted population growth curve and the real data on the same
graph. Use the same range of years as in part (d). For which value of I0 do you see the
best agreement with the real data? Again, your plot should be made using a computer
 5

and should include axis labels and an explanatory title. Plot should look reasonable,
be clear, and include axis labels and title. Best agreement is a bit subjective - could
be I0 = 200 or I0 = 500.

(5) Pharmacokinetics of ibuprofen. Recurrence equation models can be used to model a

wider range of phenomena than just population growth. One application area is pharma-
cokinetics, the study of how medications move through the human body. Suppose that we
are measuring the concentration (in mg/liter) of ibuprofen (“Advil”) in a patient’s blood
every hour. Consider:
• When the patient takes a pill, it takes an hour be absorbed in the blood.
• Each pill increases concentration of ibuprofen by 40 mg/liter.
• Ibuprofen has first-order elimination kinetics - 24.25% of ibuprofen in the blood is
eliminated each hour.
• The patient takes a pill every 6 hours. They take the first pill at t = 0.

(a) The word equation for the concentration of ibuprofen ct after t hours is
ct+1 = ct + amt. absorbed in blood − amt. eliminated in one hour
Starting with the word equation, find the concentration of ibuprofen in the blood for
the first seven hours (i.e., find c0 , c1 , c2 , ..., c7 ).
We start by noticing that c0 = 0, since the person has not taken any ibuprofen yet.
Writing down the values at each step:
c1 = 40 + (1 − 0.2425)c0 = 40
c2 = (1 − 0.2425)c1 = 30.3
c3 = (1 − 0.2425)c2 = 22.95225
c4 = (1 − 0.2425)c3 ≈ 17.3863
c5 = (1 − 0.2425)c4 ≈ 13.1701
c6 = (1 − 0.2425)c5 ≈ 9.9764
c7 = 40 + (1 − 0.2425)c6 ≈ 47.5570
Note: it’s possible some students might assume that ibuprofen undergoes elimination
kinetics during the absorption process, so that c1 = (1 − 0.2425) ∗ 40 (or equivalent,
that c0 = 40. I am fine with either strategy as long as it is self-consistent. This could
also affect their choices in part (b).
(b) Write down a recurrence equation for Cn , the concentration of ibuprofen one hour after
pill n is taken (note: not the same equation as in part a; this is a different quantity).
Cn+1 = 40 + (1 − 0.2425)6 · Cn
(c) Using your equation from part (b), compute C1 , C2 , ..., C5 .

C1 = 40 + C0 (0.7575) = 40
C2 = 40 + C1 (0.7575)6 ≈ 47.5571
C3 = 40 + C2 (0.7575)6 ≈ 48.9848
C4 = 40 + C3 (0.7575)6 ≈ 49.2546
C5 = 40 + C4 (0.7575)6 ≈ 49.3056
6

(d) You should see that Cn appears to be reaching a limiting concentration as n increases.
Find this limiting concentration by finding the fixed point of your equation in part (b).
The fixed point satisfies
C ∗ = 40 + C ∗ (0.7575)6
40
⇒ C∗ = ≈ 49.3174
1 − 0.75756

(e) We can find the limiting concentration using an alternate method! Derive an explicit
formula for Cn (hint: write down C2 , C3 , and C4 in terms of C1 , and look for a pattern).
Then take lim Cn .
n→∞
By writing down C2 , C3 , etc. in terms of C1 , and looking at the pattern (need to show
this work), the explicit formula is
k
1 − rk+1
X  
k
Ck+1 = 40 r = 40 ,
1−r
j=0

where r = 0.75756 . Because |r| < 1, we have

 
1
lim CK+1 = 40 ≈ 49.3174,
k→∞ 1−r
which matches what we got in the previous part.