Downloaded from www.clastify.

com by Discordkitten

MATHEMATICS STANDARD LEVEL INTERNAL ASSESSMENT;

A MATHEMATICAL APPROACH TO OPTIMIZING BLOOD VESSEL BRANCHING

INTRODUCTION
PERSONAL INTEREST
Coming from a family with a history of both hyper- and hypo-tension, I have always had a general interest in the workings of
the circulatory system - particularly how the circulatory system can fail. This is an interest which I’ve expressed in the
subjects I take in the IB, such as biology and physics, and has also played a role in motivating me to pursue a future
career in medicine. The moment when my interest in the circulatory system intersected with mathematics was when
learning about differential optimality in class. I remember being taken aback by the versatility of optimization and its
seemingly endless applications in the real world. This got me thinking; “could optimality principles be applied to the
circulatory system, and how significant could these applications be for human health?”. After doing some further
research into the applications of optimality in biological organisms inspired me to focus on the branching of blood
vessels, an overlooked yet essential aspect of the circulatory system’s functionality. Not only would this topic be an exciting
application of differential optimality in biology, it would also allow me to apply knowledge from my physics and biology
classes at school to view medicine from a new, mathematical perspective.

AIM OF THE EXPLORATION

In 1878, German zoologist Wilhelm Roux made a set of hypotheses regarding the angle at which blood vessels branch in the
circulatory system . These hypotheses were made on the basis of a set of real-life observations by Roux. This exploration
1
will aim to visualize Roux’s hypotheses by constructing a model of a branching blood vessel which mathematically
optimizes the branching angle. Ultimately, this mathematical model will be tested against Roux’s hypotheses in order to
determine the model’s accuracy.

BACKGROUND INFORMATION
In order to understand this exploration, an adequate understanding of the circulatory system is necessary. The circulatory
system consists of blood vessels (arteries, veins and capillaries) which transport blood around the body . In order to do this,
2
blood vessels such as arteries need to branch into smaller vessels, forming “arterial trees” which allow the blood to intercept
body tissues . The aorta, which is the largest artery, branches into smaller arteries (the arterioles), which branch into the
3
smallest blood vessels (the capillaries) . This branching pattern will be the main focus of this exploration. 45
Blood in itself is also an important topic to explain, particularly the way in which blood flows around the body. The flow
of blood is said to be “laminar” , meaning that blood flow can be imagined as consisting of many parallel layers of the fluid 5
sliding past one another, with a resistance existing between layers . This resistance is known as the “viscosity” of blood .
6 6
The opposite of laminar flow is “turbulent” flow, which is characterized by the “chaotic movement of particles in a
fluid” . 78
The nature of blood as a fluid is also important to understand. Blood is known as an incompressible, non-Newtonian fluid,
meaning that the density of blood (mass per unit volume) remains constant yet its viscosity changes depending on the amount
of force applied to it . Blood is specifically known as a “shear-thinning” liquid, meaning that blood becomes less viscous 9

1 Kurz, Haymo, Konrad Sandau, and Bodo Christ. 1997. “On the Bifurcation of Blood Vessels — Wilhelm Roux’s Doctoral Thesis (Jena 1878) — A
Seminal Work for Biophysical Modelling in Developmental Biology.” Annals of Anatomy - Anatomischer Anzeiger 179 (1): 33–36.
https://doi.org/10.1016/S0940-9602(97)80132-X.
2 Khan Academy. n.d. The circulatory system review . Accessed September 8, 2019.
https://www.khanacademy.org/science/high-school-biology/hs-human-body-systems/hs-the-circulatory-and-respiratory-systems/a/hs-the-
circulatorysystem-review.
3 Ibid.
4 Ball, Karen. 2018. Academy of Ancient Reflexology. August 15. Accessed September 15, 2019.
http://academyofancientreflexology.com/author/karen-ball/page/4/?cv=1.
5 Klabunde, Richard. 2018. Turbulent Flow. March 1. Accessed September 8, 2019. https://www.cvphysiology.com/Hemodynamics/H007.
6 What is Laminar Flow? 2018. Accessed September 8, 2019. https://www.simscale.com/docs/content/simwiki/cfd/what-is-laminar-flow.html.

https://www.princeton.edu/~gasdyn/Research/T-C_Research_Folder/Viscosity_def.html. Princeton. 2018. What is viscosity? Accessed September

8, 2019.

7 “What Is Turbulent Flow - Turbulent Flow Definition.” Nuclear Power. Accessed February 11, 2020.
https://www.nuclear-power.net/nuclear-engineering/fluid-dynamics/turbulent-flow/.
8 Hill, Kyle. 2015. What kind of liquid is blood? October 30. Accessed September 8, 2019. https://archive.nerdist.com/what-kind-of-liquid-is-
blood/. Ibid. 11 Ibid.
1
Downloaded from www.clastify.com by Discordkitten

the more force is applied it . Conversely, a fluid whose viscosity 10

remains constant independent of the forces applied to
it is called a Newtonian fluid.

ASSUMPTIONS MADE WHEN CONSTRUCTING THE MODEL

As with most mathematical models of real-life concepts, the model which will be constructed in this exploration is
accompanied by a set of simplifying assumptions. A first assumption made concerns the nature of blood as a fluid. In this
exploration, blood will be modelled as a Newtonian fluid, despite being classified as a non-Newtonian fluid in real life . 11
Given the fact that the viscosity of non-Newtonian fluids changes due to the amount of force applied to the fluid, modelling
blood as a non-Newtonian fluid would be very difficult since the force applied to blood is different in different blood
vessels, meaning that the viscosity of blood would be different in different blood vessels around the body. Therefore, in
order to apply math which is commensurate with the maths standard level course, blood will be assumed to have a
constant viscosity, as is the case with Newtonian fluids. This will, ultimately, help simplify the mathematical operations
in this exploration.
Secondly, blood will be assumed to exhibit laminar flow, as is the case in real life. Not only will this assumption allow me
to construct a model that is more accurate in real life, but it
will also allow me to explore concepts such as resistance to flow
which are important when optimizing blood flow in blood vessels.

Figure 1: visual model illustrating the branching path

of a blood vessels

A last assumption made is with regards to the structure of blood vessels. As is shown in Figure 1, blood vessels can
have different shapes and may branch in different ways. Given the complexity of blood vessel branching, this exploration
will assume that blood vessels have a constant, linear shape. This allows for an appropriate estimation of the shape of
blood vessels, despite the fact that blood vessels may also be bent. Moreover, blood vessels will be considered to be rigid in
this investigation - not flexible as they are in real life. This estimation will further simplify the mathematics used in the
investigation given that accounting for the flexible nature of blood vessels would be very complex. Another assumption
made is that blood is an incompressible fluid, as is also the case in real life. This will further contribute to the accuracy of my
investigation. Additionally, the viscous flow which is characteristic of incompressible fluids is essential for this exploration,
as described by Poiseuille’s Law (explained later on in this exploration).

The last assumption made in this exploration is that smaller blood vessels branch from larger, primary blood vessels.
Therefore, the radius of the branching blood vessel is assumed to be smaller than the radius of the primary blood vessel.

RELEVANCE OF THE EXPLORATION

The application of optimization to blood vessel branching will allow me to gain a better understanding of the circulatory
system, but also has relevance in the real world, particularly in the field of organ engineering. Organ engineering is a very
fast-advancing field, with the first organ containing blood vessels having been 3D printed in April 2019 . Since organ
12
engineering focuses on producing biological substitutes to injured or diseased organs, the ability to apply the
mathematical approach of optimality to blood vessel branching may be valuable in modelling blood vessels that minimize the
energy expended by the body for blood circulation, thus making artificial organs more efficient and thus more successful.
While this exploration won’t provide an absolute model of a perfectly optimized blood vessel (due to the many factors
2
Downloaded from www.clastify.com by Discordkitten

involved in blood flow), it may shed light on some of the approaches and difficulties in developing, for example, artificial
organs.

UNDERSTANDING POISEUILLE’S LAW

In order to construct a first blood vessel model, it’s necessary to begin by mathematically describing the flow of a fluid. One
way to do this is by utilizing Poiseuille’s Law, a physical law which was experimentally derived by French physicist Jean
Poiseuille in order to explain the laminar flow of incompressible fluids such as blood through a tube. Poiseuille’s law is 13
shown in equation (1) below: 14

(1)

When applying equation (1) in the context of blood, Q is the volumetric flow rate of blood ( m 3s-1), ΔP is the change

in pressure of blood ( Pa), r is the radius of a blood vessel ( m), μ is the viscosity of blood (Pa · s) and L is the
length of blood vessel to the fourth power (vessel ( m). As is evident from the equation, the volumetric flow rate r4) and
inversely proportional to the length of the blood vessel ( Q of blood is proportional to the radius of the blood L).

12
https://www.technologyreview.com/f/613316/a-3d-printed-heart-with-blood-vessels-has-been-made-using-human-tissue/. Jee, Charlotte. 2019. A 3D-

printed heart with blood vessels has been made using human tissue. April 16. Accessed September 8, 2019.

1314
Brittanica. 2019. Jean-Louis-Marie Poiseuille . Accessed September 8, 2019. https://www.britannica.com/biography/Jean-Louis-Marie-Poiseuille.

https://opentextbc.ca/physicstestbook2/chapter/viscosity-and-laminar-flow-poiseuilles-law/. OpenStaxCollege. 2018. Viscosity and Laminar Flow;

Poiseuille’s Law . Accessed September 8, 2019.

MODEL 1: MINIMIZING HYDRAULIC RESISTANCE DERIVING

HYDRAULIC RESISTANCE
From my physics classes, I know that a certain amount of work must be done to push a certain volume of a fluid through a
tube. The amount of work which must be done can be expressed as being equivalent to the amount of resistance in the tube;
the greater the resistance, the greater the amount of work which must be done. Ultimately, when applying this physical law to
the circulatory system, it seems that the lower the resistance to the flow of blood in a blood vessel, the smaller the amount
of work which must be done by the heart to pump that blood around the body. This, essentially, means that less
energy will need to be expended by the body in order to sustain the circulatory system, improving the system’s
efficiency and thus biologically “optimizing” it. Critically, the first mathematical model which I can construct could optimize
a branching blood vessel by minimizing the hydraulic resistance to blood flow in blood vessels.
In order to derive hydraulic resistance, it’s helpful to refer to another aspect of physics; electricity. As a physics student, I
am familiar with Ohm’s Law, an equation illustrating the relationship between voltage, current and resistance in an electrical
circuit . Ohm’s Law states that voltage 15 V is equal to current I multiplied by resistance R (V = IR ). Importantly, it
seems that Ohm’s Law can be seen as a hydraulic equivalent to Poiseuille’s Law, following the theory of the electronic-
hydraulic analogy . In this sense, voltage 16 V is comparable to the change in pressure 17ΔP , current I is comparable
to the volumetric flow rate Q, and electric resistance Re is comparable to hydraulic resistance Rh . Using this analogy, I
am able to rewrite Ohm’s Law in a hydraulic form, as seen in equation (2):

V = IR e
= ΔP = QR h
= Q = ΔRP h (2)

By rewriting the equation in this form, I saw that I am able to equate the value of Q in Poiseuille’s Law [equation
(1)] and the value of Q in Ohm’s Law [equation (2)], giving me an expression for hydraulic resistance, as seen in equation
(3):
3
Downloaded from www.clastify.com by Discordkitten

ΔP =4

= ΔRP = πΔ8μPrL 4

= R = π8μr L4 (3)

By examining equation (3), several comments regarding hydraulic resistance to blood flow can be made. Firstly, it can be
seen that only 3 variables; blood vessel length ( L), viscosity (μ) and blood vessel radius (r), affect hydraulic
resistance, as 8 andpower ( π are both constants. Secondly, the hydraulic resistance is inversely proportional to the blood

vessel radius to the fourthr4 ), meaning that the hydraulic resistance increases as the radius of the blood vessel decreases.
This makes sense, as a smaller radius means that the same volume of blood needs to be pushed
through a smaller volume, which increases the resistance to flow. Additionally, the hydraulic resistance is proportional to
the length of the blood vessel ( L), meaning that hydraulic resistance increases as the length increases. This also
makes sense as longer vessels have longer walls which provide more friction, increasing the resistance to flow. Lastly, it’s
evident that hydraulic resistance is proportional to the blood viscosity. This makes sense since viscosity is the resistive force
between the many “layers” of blood in laminar flow.

CONSTRUCTING A VISUAL MODEL

While my understanding of hydraulic resistance was strengthened by deriving equation (3), I wanted to find a way to better
contextualize how this equation would affect the flow of blood in a blood vessel. Consequently, I decided that making a
visual model of a branching blood vessel (applying the previously made assumptions in this investigation) would best allow
me to proceed in this exploration. The visual model which I created is shown in Figure 2.

15
16 AspenCore. n.d. Ohms Law and Power. Accessed September 8, 2019. https://www.electronics-tutorials.ws/dccircuits/dcp_2.html.

https://www.hydraulicspneumatics.com/technologies/hydraulic-electric-analogies-part-1. Johnson, Jack. 2014. Hydraulic-Electric Analogies, Part 1 .

April 18. Accessed September 8, 2019.

17
Ibid.
Figure 2: diagram of branching blood vessel (self-made diagram)

4
Downloaded from www.clastify.com by Discordkitten

When looking at Figure 2, it was necessary for me to consider where a branch would arise from the “primary vessel” AOC.
As mentioned earlier, the radius of the primary blood vessel r1 is assumed to be larger than the radius of the branching
vessel r2. Referring back to equation (3), I need to consider minimizing the length and maximizing the radius of the
branching path in order to minimize the hydraulic resistance to blood flow. This can be done by considering the right-angled
triangle ABC in Figure 2.
If a branch were to originate from point A (Branch 1), the total length of the branching path AB would be
minimized. This can be deduced using Pythagoras’ theorem, which states that “ a2 + b2 = c2 ” where a is one side

of the triangle (AC), b is another side of the triangle (BC) and c is the hypotenuse of the triangle (AB). Ultimately,
Pythagroas’s theorem shows that the length of Branch 1 (which is the length of the hypotenuse) will be smaller than
any other combination of length which join point A to point B. However, a branch originating from point A will also have the
effect of decreasing the average radius of the branching path AB, given that a larger proportion of the branching path
r
will be in Branch 1 which has a smaller radius ( 2) than the primary vessel.
On the other hand, if a branch were to originate at point C (Branch 2) the average radius of the branching path ACB would
be maximized, given that the larger proportion of the branching path will be in the primary vessel which has a larger radius (
r1) than the vessel branch. However, a branch originating at point C will also have the effect of increasing the total
length of the branching path, as is also evident from applying Pythagoras’ theorem.
When considering these two fictional branches, it became evident to me that the branch from the primary vessel should
originate somewhere in between points A and C at a point O where the branching path has the optimal ratio of average radius
to length in order to optimally minimize the hydraulic resistance against blood flow along the path AOB. Given this, it
becomes necessary to find at what angle θ this optimal branch should originate.

CREATING AN OPTIMAL EQUATION

In order to minimize the hydraulic resistance to blood flow, I need to employ optimality principles using differential
equations. In order to do this, I must first alter equation (3) in order to express the total hydraulic resistance in the branching
path AOB. In this model, the blood’s viscosity is constant, as blood is modelled as a Newtonian fluid, therefore making the
product of equal to a constant k. This is highlighted in the equation below:

R = k L4
r
Utilizing this equation, the total hydraulic resistance RT in the branching path AOB will incorporate the lengths and radii
of sections AO and OB : 18

( +L
RT = k Lr 141 r 24
2
) (4)

In equation (4), L1 represents the length of the blood vessel segment AO and L 2 represents the length of the blood
vessel segment OB. I decided to further develop equation (4) by trigonometrically deriving the lengths L1 and L2.
I did this by visualizing triangle OBC in a self-made diagram, shown in Diagram 1, and utilizing trigonometric principles,
particularly sine, cosine, cotangent (which is defined as ) and cosecant (which is defined as ). My derivations
of side L1 and L2 are illustrated below.

18
Adam, John. “Blood Vessel Branching: Beyond the Standard Calculus Problem.” Mathematic Magazine, June 2011.
L 1 = AO

5
Downloaded from www.clastify.com by Discordkitten

= AC − OC Diagram 1: trigonometric analysis of blood vessel branching (self-made diagram)

= AC −( BC · )
= AC − B Ccot(θ)
= L 4 − L 3 cot(θ)

L 2 = OB
= hypotenuse
hypotenuse
= opposite· opposite
= BC ·
= B Ccsc (θ)
= L 3csc(θ)

In these formulae, L 3 is the length of the blood vessel segment BC and L4 is the length of the blood vessel segment

AC. Substituting the above values of L 1 and L 2 into equation (4) creates the following equation:

R
T = k( L −4 L 3
r 14
cot(θ)
+
L 3 csc (θ)
r 24 ) (5)

In order to find the value of angle θ which minimizes the hydraulic resistance, I need to find the minimum value of
the function R T, also referred to as the optimum solution to the function. In order to do this, I can use differential
calculus to identify the turning points of this function, which may be the maximum or minimum points. First, I
must find the first derivative of equation (5). Equation (5) includes two terms which aren’t included in the maths
standard level syllabus, namely cosecant and cotangent, whose derivative are unknown to me. However, I am able to
mathematically prove that “the first derivative of csc(θ)is − csc(θ)cot(θ) and the first derivative of cot(θ) is − csc2(θ)” , as is
shown below:9

csc(θ) = cot(θ) = =
Therefore, ddθ[csc(θ)] = ddθ[sin1(θ)] Therefore, ddθ[cot(θ)] = ddθ[tan1(θ)]

Using the quotient rule: Using the quotient rule:

ddθ[ ]= =− d [tan1(θ)] = (tan(θ))(0) − (1) sin 2 (θ)

cos2 (θ)

(tan2(θ) ) = − tan2(θ) = − ( ) = − dθ

Given that csc(θ) = and cot(θ) = : Given that csc(θ) = :

ddθ[ ]=− =− · = − csc(θ)cot(θ) [ ]=− · = − csc(θ) · csc(θ) = − csc2(θ)
d
d θ

Ultimately, this knowledge allows me to find the first derivative of equation (5), as is shown below:

= k{
dRdθ T r1 14 [ddθ (L )4 − ddθ (L 3cot(θ))] + Lr 243 [ddθ (csc(θ))]}

{1
= dRdθ T = k r 14 [0 − L [3 ddθ (cot(θ))]] + Lr 243 [− csc(θ)cot(θ)]}

9 Table of Derivatives. 2005. Accessed September 15, 2019. http://www.math.com/tables/derivatives/tableof.htm.

6
Downloaded from www.clastify.com by Discordkitten

= dRdθ T = k{r1 14 [(L 3csc (θ))]2− Lr 243 [csc(θ)cot(θ)]}

−
= dRdθ T = kL {3 r1 14 [(csc (θ))]2 1
r 24 [csc(θ)cot(θ)]}
dR T csc (θ) 2
csc(θ)cot(θ)

= dθ = kL {3 r 14 − r 24 }

The minimum hydraulic resistance would occur when dR

dθ
T
= 0 or, in other words, when the gradient of the function of

equation (5) is equal to 0. Ultimately, when dR

dθ
T
= 0, I can solve to find θ min , the optimal branching angle which

minimizes hydraulic resistance:

{
= kL 3cscr (θ) 214− csc(θ)r cot24 (θ)} =0

= csc (θ)24 = csc(θ)cot4 (θ) r 1

r2

r 24 csc(θ)cot(θ)
= =
1

= (rr 21 ) 4 = csccot(θ)(θ)

1
tan(θ)
= (rr 21 ) 4 = 1
sin(θ)

= (rr 21 ) 4 = tansin(θ)(θ)

= (rr 21 ) 4 = cos(θ)
−1 r2 4
((
= θ min = cos r1 )) (6)

While equation (6) showed me a function for the optimal branching angle which minimizes the hydraulic resistance, it didn’t
give me a visual indication of how the optimal branching angle changes when the ratio of r2 to r1 changes. For this
reason, I decided to graph equation (6) using the online program Desmos. In this graph (Graph 1), the ratio r
r 2
1 was
considered as the x-value and the value of θ in radians was considered as the y-value. This creates a graph of the function

(f x) = cos−1(x4).

7
Downloaded from www.clastify.com by Discordkitten

Graph 1: graph of equation (6) with adjusted domain

Defining the domain and range of Graph 1 is important. Since the range of (f x) = cos (x4) is {− 1 ≤ cos(x) ≤ 1}, the domain
of cos −1(x) must be { − 1 ≤ x ≤ 1}. However, when considering this domain in the context of the blood vessel model, I
came to the realization that it isn’t suitable. Given that this model assumes that a smaller vessel branches from a larger
vessel, the value of rr 12 must be larger than 0 and smaller than 1 (since 0 < r2 < r1). Therefore, the suitable domain
for this graph is {0 < x < 1 }, as is illustrated in Graph 1. Moreover, while the range of (f x)
is { }, the suitable range for this graph is { 0 < cos −1(x )4 < }, given that the
branching angle must be greater than 0º (0 radians) but smaller than 90º ( radians). This is also seen in Graph 1.

The value of θ min expressed in equation (6) could either represent a local maximum or a local minimum

value, given that

the first derivative of a graph can be 0 at both of these points. Given this, it is necessary to confirm that the value for

θ min shown in equation (6) is, in fact, the minimum value. To do this, I can figure out

the second derivative of 2

R T at the point where ( =x θ ). If d R2 T > 0 , then equation

(6) is a minimum value. Conversely, if d R

2 2
T
< 0 , then equation (6) is a
maximum dθ dθ

value. I will determine the second derivative of R T by finding the derivative of the first derivative, as is shown below:

d 2R T dR T csc (θ)2 csc(θ)cot(θ)

= (kL (
dθ 2 dθ 3 r 14 − r 24 )
[ −(1 [
= kL ((3r1 14 d
d θ [csc (θ)]])2 d
r 24 d θ [csc(θ)cot(θ)]])

Using the chain rule and the product rule I was subsequently able to find the derivatives for the terms csc2(θ) and csc(θ)cot(θ)
respectively, using their derivatives which I mathematically proved previously:

= kL ((3r1 14 (− 2csc (θ)2cot(θ))) − (r1 24 [(− csc (θ))2 × csc(θ)) + (cot(θ) × (− cot(θ)csc(θ))]])
2csc (θ)2 cot(θ) −csc (θ) − 3 cot (θ)2 csc(θ)
8
Downloaded from www.clastify.com by Discordkitten

−
= kL (−3 r 14 r 24 )

Plugging in the value of θ (θ min = cos ((rr 1

2
) 4) ) from equation (6) and solving gives: −1

The above equation represents the second derivative of RT, but due to its complexity and the many variables involved in
it I was still unable to deduce whether its value was positive (denoting a minimum point on the graph) or negative (denoting
a maximum point on the graph). To overcome this problem, I decided to choose random values for the radii r1 and
r r . For example, if r = 1 and r =
r2 and plugging them in to the equation, keeping in mind that this model assumes that 1 > 2 1 2 0.5:

Given that k and L3 are always positive, the value of the second derivative is positive (16.03kL
3 > 0) , indicating that the value for θ
in equation (6) is, in fact, a minimum value.

MODEL 2: MINIMIZING THE ENERGY REQUIRED TO MAINTAIN THE BLOOD VESSEL STRUCTURE
After constructing my first model [equation (6)], which minimized the hydraulic resistance to flow, I couldn’t help but
wonder if there were other factors which could be optimized in order to maximize the efficiency of the circulatory system.
After doing some research I came across one such factor; the energy required to maintain the blood vessel structure. In
deriving equation (5), I assumed that the only energy which an organism must expend is in the pumping of blood
around the body. However, it is also suggested that there is a certain amount of energy which an organism must expend
to maintain the structure of a blood vessel (i.e. keeping the blood vessel dilated). This energy expenditure must also be
minimized in order to maximize the efficiency of the circulatory system. Theoretical biologist Robert Rosen proposed this
form of energy expenditure by the body, stating that organic structures carry with them a “metabolic cost which roughly
represents the energy expenditure required by an organism to maintain said structure” . Ultimately, if this energy
expenditure is10 minimized, then an organism would need to expend less energy to maintain the structure of blood
vessels, thus optimizing the circulatory system. Rosen suggested that the energy required to maintain a given anatomical
structure would be proportional to the volume of that structure. Given this, “it can be assumed that the energy required
to maintain a blood vessel structure, S, would be proportional to the volume of the blood vessel, which can be
assumed to be a cylinder. In this equation, Lπr represents the volume of a blood vessel, with
2

L being the length of the blood vessel and r

being the radius of the blood vessel” : 21

S = B(Lπr )2

10 Rosen, Robert. 1967. Optimality Principles in Biology. http://link.springer.com/openurl?genre=book&isbn=978-1-4899-6207-2.

9
Downloaded from www.clastify.com by Discordkitten

The above equation, suggested by Rosen, included a constant B , whose

significance wasn’t specified. In order to determine
what this constantwork done by a fluid against a certain amount of external pressure” . The equation for pressure-volume
work is shown B represents, I once again turned to my knowledge in physics, particularly to pressure-volume work; “the

22 below, where P is pressure in Pascals and V is volume in m3 ;

"pressure − volume" work = − PV

In physics, “work” is the transfer of energy from one point to another, meaning that the equation for pressure-volume
could express the energy required to maintain the blood vessel structure in the body. Ultimately, this would suggest that
the constant B in Rosen’s equation is equal to -P (the external pressure on a blood vessel), given that “ Lπr 2 ”
represents the volume of a blood vessel V. Ultimately, applying this knowledge allows me to alter Rosen’s equation to
express the total energy, ST, required to maintain the structure of the branching path AOB :

S T = B(L π1 r 21 + L π2 r )22

I then substituted the values of L1 and L2 which I trigonometrically derived earlier into this equation:

S T = B{[(L 3csc(θ))πr ]12 + [(L 4 − L 3cot(θ))πr ]} 22 (7)

Proceeding just as in the previous model, I proceeded to find the optimum solution of equation (7) by determining
its first derivative, as shown below.
dSdθ T = B(L π3 r 21 · ddθ csc(θ)) + (πr (22 ddθ (− L 3cot(θ)) + ddθ L ) 4

= B(L π3 r 2
1 · (− cot(θ)csc(θ))) + (πr ((−22 L 3 · ddθ cot(θ)) + 0)

= B − (L π3 r 2
1 · (cot(θ)csc(θ))) + (− L π3 r ((−22 csc (θ)))2
= B(− L π3 r 21cot(θ)csc(θ) + L π3 r 22csc (θ))2
= BL π3 csc(θ)[− r 21cot(θ) + r 22csc(θ)]

The critical point of this function would occur when dS

dθ
T
= 0. I can therefore solve to find θmin, the optimal branching
angle which minimizes the energy required to maintain the blood vessel branching structure:

BL π3 csc(θ)[− r 21cot(θ) + r 22csc(θ)] = 0

= (r 22csc(θ)) = (r 21cot(θ))

= =
cot(θ)
r2 csc(θ)
cos(θ) r1
sin(θ)
1 r2 2
sin(θ)
== ( )
r1

= cos(θ) = (r 2 ) 2
r1

−1 r2 2
10
Downloaded from www.clastify.com by Discordkitten

((
= θ min = cos r1 )) (8)
I also decided to graph equation (8) using Desmos in order to better visualize it. In this graph (Graph 2), I considered the
ratio rr 12 as the x-value and the value of θ in radians as the y-value. This creates a graph of the function f(x) = cos−1(x2).

21
22 Rosen, Robert. 1967. Optimality Principles in Biology. http://link.springer.com/openurl?genre=book&isbn=978-1-4899-6207-2.
Khan Academy. n.d. Pressure-volume work . Accessed October 6, 2019.
https://www.khanacademy.org/science/chemistry/thermodynamics-chemistry/internal-energy-sal/a/pressure-volume-work.
Graph 2: graph of equation (8) with adjusted domain
Evidently, Graph 2 is different than Graph 1. Although both graphs incorporate the same ratio

r 2
r 1 on the x-axis, in Graph 1

this ratio is graphed to the fourth power while in Graph 2 it is graphed to the second power. Despite this difference, the
appropriate domain and range of Graphs 1 and 2 remains the same (the domain is {0 < x < 1 } and the range is {0 <
cos −1(x )2 < }), as both equations are inverse cosine equations and both abide to the same assumptions made in this
exploration. Again, in order to confirm that the value for θ expressed in equation (8) is, in fact, the minimum value, I
will figure out its second derivative.

d 2S T dS T 2 2
dθ 2 = dθ {BL π3 csc(θ)[r 1csc(θ) − r 2cot(θ)}

= BL π[((−3cot(θ)csc(θ))(r 21csc(θ) − r 22cot(θ))) + (csc(θ))(r 21 · ddθ (csc(θ)) − r 22 · ddθ (cot(θ)))]

= BL π[((−3cot(θ)csc(θ))(r 21csc(θ) − r 22cot(θ))) + (csc(θ))(r 21 · (− cot(θ)csc(θ)) − r 22 · (− csc2(θ)))]

= BKL π3 csc(θ)[(r 22csc (θ))2 + (2r 21cot(θ)csc(θ)) + (r 22cot (θ))]2

Plugging in the value of θ (θ min = cos −1((rr 2

1 ) 2) ) from equation (8) and solving gives:

= BL π3 csc(cos
dd 2θ S2 T −1((rr 12 ) )[(2r 22csc (2 cos −1((rr 12 ) 2)) + (2r 12cot(cos −1((rr 12 ) 2)csc(cos −1((rr 12 ) 2)) + (r 22cot (2 cos −1((rr
12 ) ))]2
11
Downloaded from www.clastify.com by Discordkitten

Once again, I used the same random values for radii r 1 and
r 2 , r 1 = 1 and r 2= 0.5, to determine whether equation (8) optimizes

function S T and is a minimum solution to it:

d
d 2θ
S
2
T −1
((0.5) )[((0.5)2 2csc (2 cos −1((0.5) ))2 + (2cot(cos −1((0.5) )2 csc(cos −1((0.5) ))2 + ((0.5)2cot (2 cos −1((0.5) ))]2
= BL π3 csc(cos

= BL π3 csc(1.318)[((0.25)csc (1.318))2 + (2cot(1.318)csc(1.318)) + ((0.25)cot (1.318))]2

= BL π(1.033)(0.267)3+ ((0.534) + (0.017))

= 0.276BL π3 + 0.551

Given that B, L 3 and π are always positive, the value for the second derivative is greater than zero
0.276( BL π3 + 0.551 > 0). This means that the value
for θ in equation (8) does, in fact, express a minimum value.

MODEL 3: FINAL OPTIMAL MODEL

After constructing models 1 and 2, I wanted to develop a final optimal equation which considered minimizing the hydraulic
resistance and the energy required to maintain the blood vessel structure. I was able to do this by combining the equation for
RT and the equation for S T which I had developed earlier. My newly formed equation, O T, is shown below:

OT = R T + S T
= k(Lr 141 + Lr 242 ) + B(L 1r 21 + L 2r )22

I further simplified this equation by removing the common term L 1 and rearranging. For simplicity, the terms (rk 41 + Br )12 and

(rk 42 + Br )22 can be named W 1 and W 2 respectively:

= L (1 rk 14 + Br )21 + L (2 rk 24 + Br )22

= L 1W 1 + L 2W 2

I then proceeded to substitute the values of L 1 and L 2 derived earlier into the above equation:

= (L 4 − L 3cot(θ))(W )1 + (L 3csc(θ))(W )2
= W (1 L 4 − L 3cot(θ)) + W 2L 3csc(θ)

Proceeding just as in the previous models, I found the optimum solution of this equation by determining its first derivative:
dOd T = W [dd (L )4 − L 3 ddθ (cot(θ))] + W 2L 3 ddθ [csc(θ)] θ 1 θ

= W [01 − L (−3 csc (θ))]2 − W 2L 3cot(θ)csc(θ)

= W 1L 3csc (θ)2 − W 2L 3cot(θ)csc(θ)
= L 3csc(θ)[W 1csc(θ) − W 2cot(θ)]

= L 3csc(θ)[W (2 WW 21 csc(θ) − cot(θ))]

12
Downloaded from www.clastify.com by Discordkitten

Here again, I further simplified the equation by removing the term csc(θ) from the parentheses:

= W 2L 3csc(θ)[csc(θ)(WW 21 − csccot(θ)(θ))]
cos(θ)
sin(θ)
= W 2L 3csc (θ)[2 WW 21 − 1 ]
sin(θ)

= W 2L 3csc (θ)[2 WW 21 − cos(θ)]

The critical point of this function would occur when dOdθ T = 0. At this point θ would be the optimal branching angle:

= W 2L 3csc (θ)[2 W 1
W 2 − cos(θ)] = 0 =

[WW 21 − cos(θ)] = 0

= cos(θ) = WW 21

= θ = cos −1(WW 21 )

After finding this equation, I was able to substitute the values for W 1 and W 2 from before : 11

rk4 +Br 12

= θ = cos −1( k 1 2 )
r4 + Br 2
2

−1 r2 4 k+Br 16
[((
= θ = cos ) )( 6 )] (9)

r1 k+Br 2

The value of θ in equation (9) depends on the relative values of k and B. There exist two limiting situations where the value
of θ is identical to the value of θ in the previous models: the first limiting situation is when the value of r2
approaches the value of r1 (r2 → r 1 ) . In this situation, the value of θ in equation (9) becomes identical with the value of θ
in equation (6) as:
lim k+Br 61 = 1 , thus making θ = cos −1(( r 2 ) 4) r → 2 r 1 k+Br 62 r 1

The second limiting situation is when the value of r2 approaches zero (r

2 → 0). In this situation, the value of θ in equation (9)
becomes identical with the value of θ in equations (6) and (8) as:

lim r 2 −1
(0) =
= 0 , thus making θ = cos
r → 02 r 1

11 Adam, John. “Blood Vessel Branching: Beyond the Standard Calculus Problem.” Mathematic Magazine, June 2011.
13
Downloaded from www.clastify.com by Discordkitten

However, between these two limiting situations equations (6), (8) and (9) do not agree. One way in which I attempted to
better understand equation (9) was by graphing it but, unlike equations (6) and (8), equation (9) has several different
variables and is difficult to graph. Instead, in order to better understand equation (9) I decided to better understand the
k+Br significance of k+Br 62 in the equation. Given that all four values in
6
1 this expression; 6 k, B, r 1 and r2,
are positive, the value of
k+Br 6 k+Br 1
k+Br 2 will also be positive. Essentially, the effect of the expression
k+Br 62 is to horizontally stretch the graph of

θ = cos −1((rr 2
1 ) 4) by the scale factor . If the value of
k + Br 62
k + Br 61
k + Br
k + Br
61
62 is greater than 1, it
moves the points of θ = cos (( −1
) 4)
r 2
r 1

k+Br 61 k+Br 6
1 closer to the y-axis. Conversely, if the value of k+Br 62 is
greater than zero but smaller than 1 (0 < k+Br 62 < 1) , it moves the points of θ = cos −1((rr 1
2
) 4) further away from the y-
axis.6
k+Br 1
After understanding the significance k+Br 62 in the equation, I decided to try graph equation (9) by calculating
values of θ for different values of
r1 and r2 and plotting these on a graph with x-axis “ 6 r
r
1
2
”

and y-axis “ θ”, as I had done with equations

(6) and (8). In the expression k + Br

k + Br
1
62 , the value of constant k is , where µ
is the viscosity of blood in pascal-seconds (

Pa · s). Since this value is constant it can be calculated, taking the viscosity of blood as 2.78 × 10 −3
Pa · s at body
temperature (37ºC) : 12
8 · (2.78×10 −3) k
== 3.1415926535 =
0.00708 Pa · s

I was also able to deduce the value of B in equation (9) as it represents the external pressure on blood vessels in the
body, as was explained when deriving model 2. This external pressure is 13,000 Pascals (Pa )25. When

graphing equation (6) and (8), the graph produced illustrated the values of θ at every possible ratio of r
r
1
2
.

However, to graph equation (9), I decided to apply the equation to real values of blood vessel radii, in order to make this
optimal model more applicable to real life. I was able to do this by utilizing data collected by Zamir and Phipps
which showed the diameter of arteries in the arterial tree of dogs , as illustrated in Table 1. These values are in metres
given that Poiseuille’s Law also uses values in metres. 13

Table 1: radii of different types of blood vessels in a dog’s arterial tree

Blood vessel type Diameter of blood vessel (metres/m) Radius of blood vessel (metres/m)

Aorta 10.000 5.000

Large artery 3.000 1.500

Main branch 1.000 0.500

12 Anton Paar GmbH. n.d. Viscosity of Whole Blood . Accessed September 8, 2019. https://wiki.anton-paar.com/en/whole-blood/.

https://www.webmd.com/hypertension-high-blood-pressure/qa/what-is-normal-blood-pressure. 25 Jaliman, Debra. 2017. What is "normal" blood pressure?

July 11. Accessed October 6, 2019.

13 6https://doi.org/10.1016/0021-9290(88)90188-1. Zamir, M., and S. Phipps. 1988. “Network Analysis of an Arterial Tree.” Journal of Biomechanics
21 (1): 25–34.
14
Downloaded from www.clastify.com by Discordkitten

Secondary branch 0.600 0.300

Tertiary branch 0.140 0.070

Terminal artery 0.050 0.025

Terminal branch 0.030 0.015

Arterioles 0.020 0.010

Using these blood vessel radii values, I was able to find the corresponding values of θ in radians using equation (9)

in different branching scenarios (different ratios of rr 1

2
) . These calculations can be seen in Table 2 below:

Table 2: radii of different types of blood vessels in a dog’s arterial tree

/r
Branching scenario r2 1 θ in radians
Large artery branching from the aorta 0.300 1.542

Main branch branching from the larger artery 0.333 1.534

Secondary branch branching from the main branch 0.600 1.378

Tertiary branch branching from the secondary branch 0.233 1.555

Terminal artery branching from the tertiary branch 0.835 0.991

Arteriole branching from the terminal branch 0.667 1.293

Secondary branch branching from the large artery 0.200 1.569

Tertiary branch branching from the large artery 0.047 1.571

Ultimately, I was able to plot the values calculated in Table 2 to help me visualize what the graph of equation (9) would look
like as well as how its relative shape compares to the graphs of equations (6) and (8). This is shown in Graph 3
below, where the blue dotted line is the graph of equation (6), the red dotted line is the graph of equation (8) and the
black dots are the plotted values from equation (9):

Graph 3:​​comparison of equations (6), (8) and (9) as graphs

15
Downloaded from www.clastify.com by Discordkitten

COMPARING THE MODEL TO ROUX’S HYPOTHESES

To verify whether the values of θ from equation (9) are applicable to the real world, I referred to the hypotheses
which Roux made regarding the branching of blood vessels in the circulatory system. These hypotheses are: “if one of the
two branches is smaller than the other, the primary vessel will make a larger angle with the smaller branch than
it does with the larger branch” and “branches which are so small that they don’t seem to weaken or diminish the blood
flow from the primary vessel come off from it at a large angle from about 70º to 90º” . Values for the ratio 14 r 2
r 1 from
Table 2 can be used to test these hypotheses. This requires me to convert the radian values for θ from radians to
degrees, which is done by multiplying the radian value by . As is seen, if r
r 2
1 = 0.200, then θ = 1.569 radians
(≈ 89.50º); if r
r
1
2
= 0.600, then θ = 1.378 radians (≈
78.95º); and if r
r
1
2
= 0.835 , then θ = 0.991 (≈ 56.78º) . Evidently, as the ratio of r
r
1
2
increases (and thus the size of the
branching vessel increases), the branching angle θ decreases. This adheres to Roux’s hypotheses, which state that smaller

branching vessels come off from the primary vessel at larger angles, thus showing how my model’s results are in line
with Roux’s hypotheses.

EVALUATION OF THE MODELS

In order to evaluate the models I developed, it's necessary to consider the assumptions made in this exploration and how
these assumptions either improve or weaken the model’s application to the real world. One of the more prominent
assumptions made was in the modelling of blood as a Newtonian fluid. Although this assumption greatly simplified the
mathematics required to construct the models, it does hinder the real-life applications of these models to the circulatory
system, such as in capillaries. The diameter of capillaries is roughly that of singular red blood cells , meaning red blood
cells 15
can only travel through capillaries in a single layer. Given this, it is unreasonable to assume that blood travels as
layers of fluid sliding past each other with viscosity in capillaries, as the present models assume. This presents the model’s
limitations in applications in, for example, a medical context. Another limiting assumption made is that the blood vessels are
perfectly linear, cylindrical and rigid. Although this assumption simplified the mathematics used, it is contrary to
observations in real life, where blood vessels are often curved and made from flexible muscle fibers. This is said to have an
impact on blood dynamics, a factor which was critically overlooked in this exploration which ultimately also limits the
applications of these models to real life. Ultimately, the assumption made in this exploration allowed me to explore this topic
with a level of mathematics that is commensurate with the standard level course. If this topic were to be explored more
seriously, more advanced mathematics could be used to make the models more applicable to reality - an aspect which limited
the applicability of my own exploration.
One more aspect of these models to consider is their scope and thus the extent to which they are able to fulfill the aim of
this exploration. This exploration has been successful in producing an optimal equation which maximizes the efficiency
of the circulatory system by optimizing the vessel branching angle whilst considering two factors to be minimized; hydraulic
resistance and the metabolic cost of maintaining the blood vessel structure. Despite this success, deducing the optimal
branching angle is only a small portion of the investigation into maximizing the efficiency of the circulatory system.
For instance, while blood vessel branching was covered, blood vessel bifurcations, which arguably occur as frequently
as blood vessel branching , were not considered at all. The models focused on optimizing blood flow in the path AOB yet
overlooked 29 the significance of section C; the continuation of the primary vessel. This is an example of an extension to this
exploration. Apart from this, further extensions which could be made to this investigation would be to minimize other
factors which may increase resistance to blood flow. Returning to equation (1), it might be worth investigating the
impact which different blood viscosities have on resistance to blood flow. This would allow the model to cover a larger
scope, thus maximizing the efficiency of the circulatory system to a greater extent.
While the limitations of the models might seem overpowering, it’s also important to consider the successes of the
models created. While they aren’t perfect, the models present a commensurate and mathematically accurate way to
optimize the branching angle in blood vessels. The final optimal model was shown to adhere to Roux’s observations,
thus somewhat confirming the credibility of equation (9). Furthermore, it’s also important to consider the inherent
14 Passariello, Alessandra. "Wilhelm Roux (1850-1924) and blood vessel branching." Journal of Theoretical and Applied Vascular Research 1, no. 1
(2016): 25-40.

15 8https://www.thoughtco.com/capillary-anatomy-373239. Bailey, Regina. 2019. Understanding Capillary Fluid Exchange . August 19. Accessed
September 14, 2019.

https://training.seer.cancer.gov/anatomy/cardiovascular/blood/classification.html. 29 National Cancer Institute. n.d. Classification & Structure of Blood

Vessels . Accessed September 14, 2019.

16
Downloaded from www.clastify.com by Discordkitten

difficulties of applying optimality principles to biological organisms. This fact became increasingly apparent to me as I
worked on this exploration and understood the complexity of biological organisms. This brought me back to one of the first
questions I asked myself when beginning this exploration; “could optimality principles be applied to the circulatory
system?”. It also sparked further questions which I asked myself, such as “why should I expect optimality principles to
manifest in the natural world?”. This exploration has allowed me to discover that it is difficult to pragmatically justify
the application of mathematical principles in biology, particularly given the many factors which affect biological
systems such as the circulatory system. Ultimately, this has not only given me a better understanding of some of the
fields in which mathematics still needs to grow but has also given me hope that mathematics will one day allow us to
optimize functions of the human body; positively impacting human health and well-being.

CONCLUSION AND EXTENSIONS TO THE EXPLORATION

Although this mathematical exploration has limitations which prevent it from becoming directly applicable to real life,
the results of the exploration should not be overlooked. The many successes of this model, alongside the difficulty of
mathematically modelling biological systems, are still a step forward in understanding the applications of mathematical
principles in the natural world. I have also come to realize that the significance of this exploration stems farther than the
circulatory system. With certain modifications, e.g. to viscosity, this equation can also be applied to the transport of other
bodily fluids in vessels through the body, such as lymphatic fluid in lymph vessels.
Further research and alternative approaches to this investigation would be valuable in elevating the success of the models
created. Apart from considering other parts of the circulatory system, such as blood vessel bifurcation, it might also be
sensible to further apply the models to real-life data. For example, integral calculus can be used to find the exact length
of real-life blood vessels from a photo, which could be applied to the models to more closely mimic the true structure of the
circulatory system (instead of the original assumption that blood vessels are linear structures).

LITERATURE
Adam, John. “Blood Vessel Branching: Beyond the Standard Calculus Problem.” Mathematic Magazine, June 2011.
Anton Paar GmbH. n.d. Viscosity of Whole Blood . Accessed September 8, 2019. https://wiki.anton-paar.com/en/whole-
blood/.
AspenCore. n.d. Ohms Law and Power. Accessed September 8, 2019.
https://www.electronics-tutorials.ws/dccircuits/dcp_2.html.
Bailey, Regina. 2019. Understanding Capillary Fluid Exchange . August 19. Accessed September 14, 2019.
https://www.thoughtco.com/capillary-anatomy-373239.
Ball, Karen. 2018. Academy of Ancient Reflexology. August 15. Accessed September 15, 2019.
http://academyofancientreflexology.com/author/karen-ball/page/4/?cv=1.
Britannica. 2019. Jean-Louis-Marie Poiseuille . Accessed September 8, 2019.
https://www.britannica.com/biography/Jean-Louis-Marie-Poiseuille.
Hill, Kyle. 2015. What kind of liquid is blood? October 30. Accessed September 8, 2019.
https://archive.nerdist.com/what-kind-of-liquid-is-blood/.
Jaliman, Debra. 2017. What is "normal" blood pressure? July 11. Accessed
October 6, 2019. https://www.webmd.com/hypertension-high-blood-pressure/qa/what-is-normal-blood-
pressure. Jayanti, Sreenivas. 2018. Computational Fluid Dynamics for Engineers and Scientists.
Jee, Charlotte. 2019. A 3D-printed heart with blood vessels has been made using human tissue. April 16. Accessed
September 8, 2019.
https://www.technologyreview.com/f/613316/a-3d-printed-heart-with-blood-vessels-has-been-made-using-human-tissue/.
Johnson, Jack. 2014. Hydraulic-Electric Analogies, Part 1 . April 18. Accessed September 8, 2019.
https://www.hydraulicspneumatics.com/technologies/hydraulic-electric-analogies-part-1. Khan
Academy. n.d. Pressure-volume work . Accessed October 6, 2019.
https://www.khanacademy.org/science/chemistry/thermodynamics-chemistry/internal-energy-sal/a/pressure-volume-work.
Khan Academy. n.d. The circulatory system review . Accessed September 8, 2019.
https://www.khanacademy.org/science/high-school-biology/hs-human-body-systems/hs-the-circulatory-and-respiratory-syste
ms/a/hs-the-circulatory-system-review.

17
Downloaded from www.clastify.com by Discordkitten

Klabunde, Richard. 2018. Turbulent Flow. March 1. Accessed September 8, 2019.

https://www.cvphysiology.com/Hemodynamics/H007.
Kurz, Haymo, Konrad Sandau, and Bodo Christ. 1997. “On the Bifurcation of Blood Vessels — Wilhelm Roux’s Doctoral
Thesis (Jena 1878) — A Seminal Work for Biophysical Modelling in Developmental Biology.” Annals of Anatomy -
Anatomischer Anzeiger 179 (1): 33–36. https://doi.org/10.1016/S0940-9602(97)80132-X.
National Cancer Institute. n.d. Classification & Structure of Blood Vessels . Accessed September 14, 2019.
https://training.seer.cancer.gov/anatomy/cardiovascular/blood/classification.html.
OpenStaxCollege. 2018. Viscosity and Laminar Flow; Poiseuille’s Law . Accessed September 8, 2019.
https://opentextbc.ca/physicstestbook2/chapter/viscosity-and-laminar-flow-poiseuilles-law/.
Passariello, Alessandra. "Wilhelm Roux (1850-1924) and blood vessel branching." Journal of Theoretical and Applied
Vascular Research 1, no. 1 (2016): 25-40.
Princeton. 2018. What is viscosity? Accessed September 8, 2019.
https://www.princeton.edu/~gasdyn/Research/T-C_Research_Folder/Viscosity_def.html. Rosen,
Robert. 1967. Optimality Principles in Biology.
http://link.springer.com/openurl?genre=book&isbn=978-1-4899-6207-2.
Table of Derivatives. 2005. Accessed September 15, 2019. http://www.math.com/tables/derivatives/tableof.htm.
What is Laminar Flow? 2018. Accessed September 8, 2019.
https://www.simscale.com/docs/content/simwiki/cfd/what-is-laminar-flow.html.
“What Is Turbulent Flow - Turbulent Flow Definition.” Nuclear Power. Accessed February 11, 2020.
https://www.nuclear-power.net/nuclear-engineering/fluid-dynamics/turbulent-flow/.
Zamir, M., and S. Phipps. 1988. “Network Analysis of an Arterial Tree.” Journal of Biomechanics 21 (1): 25–34.
https://doi.org/10.1016/0021-9290(88)90188-1. Images:
Figure 1: Schreiner, Karch, Neumann, Neumann, Rodler, and End. Visual Representation of 3-Dimensional Staged
Growth Modeling a Coronary Arterial Bed. . Photograph. Vienna: International Journal of Bioelectromagnetism. Accessed
February 13, 2020. http://www.ijbem.org/volume2/number2/139-143.htm.

18